Take a deep breath

Thousands of people have been diagnosed with a disease that may not even exist. Duncan Campbell reports

EVER since the First World War and the horrors of the trenches, physicians, psychologists and psychiatrists have puzzled over how apparently healthy people can end up bedridden, breathless and bothered by heart irregularities. This condition has been given many names, including soldier’s heart, effort syndrome and neurasthenia. Patients suffer a wide variety of symptoms and no physical cause is obvious, so there is still no agreed definition or treatment.

But for forty years, one diagnosis has been in the ascendant—chronic hyperventilation, or hyperventilation syndrome (HVS). A sufferer of this hypothetical condition is believed to breathe excessively over long periods. Persistent overbreathing removes too much carbon dioxide from the blood. This increases blood pH and, says the HVS theory, triggers a combination of symptoms that can include wind, diarrhoea, muscle pains, fatigue, heart arrhythmia and anxiety.

HVS grew in prominence in the 1970s. Claude Lum, formerly a chest physician at Papworth Hospital in Cambridge, claimed that up to 10 per cent of people visiting outpatient clinics suffered from the syndrome. He also stressed that HVS was not a product of other disease, but a “primary disorder”.

In the mid-1980s, this torch was picked up by Peter Nixon, a former NHS cardiologist and prominent Harley Street consultant, who believed that HVS was a prime cause of cardiac disorders, including heart attacks. Nixon has written more than 70 journal and newspaper articles blaming HVS for everything from chronic fatigue to Gulf War syndrome. But Nixon’s work on hyperventilation is now discredited and the very existence of HVS is in doubt.

His reputation collapsed in May after he unsuccessfully sued for libel over allegations in a television documentary that he had rigged hyperventilation tests. The court case revealed that Nixon’s research and diagnostic techniques were, at best, scientifically and ethically dubious. After five days of cross-examination at the High Court in London, he withdrew his case, agreed to pay costs of £765 000, and to retire immediately from medical practice.

In a bizarre twist, two prominent researchers who supported Nixon provided the ammunition for a surprise attack in court on recent studies which appear to show that HVS does not exist. The nature of the clash suggests that some researchers have placed their faith in HVS ahead of empirical science. These events have confirmed long-standing worries among some specialists that HVS was being written about in journals and diagnosed by doctors despite a lack of rigorous scientific testing.

Even New Scientist carried an uncritical report of Lum’s and Nixon’s claims (“The hazards of heavy breathing”, 3 December 1988, p 48). One of the chief critics of HVS is William Gardner, senior lecturer in the department of thoracic medicine at King’s College School of Medicine and Dentistry in London. There is a “dearth of data based on modern quantitative criteria”, he wrote last year in Chest (vol 109, p 516).

Dangerous diagnosis

“Much research in this area is bedevilled...by the presentation of scientifically unsound data lacking in rigorous quantitative proof and with perpetuation of circular arguments,” Gardner has argued. “Not only is the term hyperventilation syndrome of dubious value clinically, its use can be dangerous in that it distracts from seeking the true causes of the increased respiratory drive.” So patients could be put at risk when HVS is diagnosed as the cause of serious disease, while the real underlying problem goes untreated.

Back in 1979, in a bid to place HVS on a sound footing, two Dutch military physicians reported that they had developed an objective diagnostic test. Their hyperventilation provocation test (HVPT) uses a capnograph, which measures the partial pressure of carbon dioxide in exhaled breath. The patient is asked to hyperventilate for 3 minutes, which can halve the level of CO2 in blood. If the pressure of CO2 in the breath does not then recover to at least two-thirds of normal after 3 minutes or so the diagnosis is HVS. But was this test valid, or could the stress of the test induce positive results? To answer this would take a comparison with the best available test. The “gold standard” would be to directly measure the level of CO2 dissolved in arterial blood and show it to be lower than normal in HVS patients.

Until recently, however, such checks could only be done in a lab, and were not performed by many physicians. In the meantime, some researchers who used the HVPT failed to obtain clear-cut results. They complained that it wrongly identified non-sufferers as having HVS—it has a high false-positive rate. Nixon advanced a different view. The real problem with the HVPT, he said, is too many false negatives. It is an “insensitive technique which causes the rejection of many patients”, he wrote in the Journal of the Royal Society of Medicine (vol 81, p 277). So, in the same paper, published in May 1988, he described an alternative—the “think test”. He connected patients to a capnograph and asked them to think about stressful events in their life. If the pressure of CO2 in their exhaled breath fell by an arbitrary 10 millimetres of mercury, Nixon diagnosed HVS. The think test produced twice as many diagnoses as the HVPT. This, he argued, suggested it was the better test.

Laboratory workbooks that Nixon disclosed in court show that, soon after inventing the think test, he had manipulated the procedure and achieved even higher rates of diagnosis. In the May 1988 paper, Nixon stressed that no mention of breathing should be made during the test, to avoid inducing the type of response the doctor was looking for. Mentioning breathing, he told me in an interview in 1991, meant that the results of the think test “didn’t prove anything”. But his workbooks show that from January 1987 he had systematically prompted patients to “let your body show me how it was breathing” as part of his think test.

The nature of Nixon’s actual clinical procedures surfaced in the television documentary, which I produced, and which was shown in 1994 on Britain’s Channel 4. In the programme, we secretly filmed Nixon being consulted by an AIDS patient, Ian Hughes, who died last year. During the test, Nixon told Hughes to “show me how you breathe when you’re angry”. Then, after recording the dip in partial pressure of CO2 from Hughes’s breath in response to this request, Nixon announced his findings. “It’s not some bloody virus [HIV] that pulls you down, but the state you get into.”

Nixon recorded his diagnosis as gross hyperventilation, and prescribed valium and an antihistamine, followed by courses of physiotherapy and counselling to be purchased at his Harley Street clinic. In court, under cross-examination, it became glaringly clear that Nixon had a poor grasp of scientific principles—even when it came to his think test. An analysis of his work by Kenneth MacRae, reader in medical statistics at the Charing Cross and Westminster Medical School in London, revealed that Nixon’s methods had the primary effect of increasing the number of false positive diagnoses.

Gold standard

There were other problems too. In his May 1988 paper, he reported on 54 patients, 16 of whom were positive on the HVPT, and 33 of whom had a positive think test. But his test missed 9 of the 16 patients who were positive on the HVPT. So the two tests seemed to be identifying different groups, making comparison meaningless. In claiming to have a better test, Nixon had also failed to compare his results with a gold standard test.

In court, Nixon was presented with MacRae’s analysis and asked: “You had no honest grounds for making the claims you did about the efficacy of the think test?” He said yes. “You agree with me?” the barrister asked. “Yes,” replied Nixon.

Worse followed. Nixon said he had not prepared protocols for his scientific studies or, if he had, he hadn’t written them down. He admitted that at Charing Cross Hospital and in private practice he experimented on patients without ethical approval. He did not prepare informed consent forms or warn subjects about the potential dangers, even though the tests carry the risk of inducing heart attacks. Nixon’s workbooks listed more than 4200 patients that he had tested since 1986, but he couldn’t say which patients had been included in which studies.

He also admitted reporting the same data differently in different places. In a paper in the Journal of the Royal Society of Medicine (vol 79, p 76), he said that 19 out of 27 patients tested for HVS using hypnosis were positive. Later, in the American Journal of Clinical Hypnosis (vol 30, p 296), he omitted three of the negative patients, improving his results to 19 out of 24. This manoeuvre, he admitted in court, “looked rather suspicious”. After his fifth day of giving evidence, the judge— who had earlier told him that he appeared to have been, at best, medically negligent—urged him to consider giving up his case, which he did.

During the court case, a subplot began to unfold involving researchers with opposing views of HVS. In July last year, a group in Amsterdam published results of a double-blind, placebo-controlled trial designed to investigate the validity of the HVPT and the hyperventilation syndrome (The Lancet, vol 348, p 154). Hellen Hornsveld and her colleagues put 115 people with suspected HVS through two provocation tests each, and asked them to list the symptoms they felt. One of the two tests was a placebo, however, in which a team member fed CO2 to the patient at a rate that kept the partial pressure in their exhaled breath constant—so the pH of the blood should not have changed.

The researchers also tested 40 healthy controls. They divided the subjects into four groups: true positives, who reported symptoms on the HVPT but not on the placebo test; false positives, who suffered symptoms during both tests; a group that felt no symptoms during the HVPT; and the healthy controls. Using standard test criteria, Hornsveld’s team found that the HVPT identified 85 people as suffering HVS. But 56 of them were false positives. And that’s not all. The researchers fitted 15 true positives and 15 false positives with small sensors that sit on the skin and constantly measure the pressure of CO2 in the blood. For two days the subjects went about their business and recorded the time of any attack of symptoms.

The results show that only a few attacks were accompanied by hyperventilation. In these cases, changes in breathing came after the onset of symptoms, so hyperventilation appeared to be the result, not the cause. The team concludes that the notion of HVS has “become untenable” and that the HVPT is “invalid as a diagnostic test”. It’s time to say “farewell to the hyperventilation syndrome”, says Hornsveld.

During the court case, one of the Dutch team, pulmonary physician Paul van Spiegel, travelled to London to interpret Nixon’s capnograph records of Ian Hughes, and to explain to the court why these and other records did not support Nixon’s diagnoses of HVS. Van Spiegel is a prominent member of a group of scientific specialists known as the International Society for the Advancement of Respiratory Psychophysiology (ISARP). At the same time, two of ISARP’s top officials had sided with Nixon. Ronald Ley, professor of psychology at the State University of New York, Albany, and Beverly Timmons, a research fellow at St Bartholomew’s hospital in London, are the two most recent presidents of ISARP.

Without consulting the society’s board, they agreed in 1994 that as soon as Nixon’s case had been resolved, he should receive the society’s Award for Distinguished Contribution to Respiratory Psychophysiology. Timmons wrote a letter, which Nixon produced in court, that described Nixon as “the obvious and most appropriate candidate” for the award. In April this year, the two officials devoted a large part of the spring issue of the society’s newsletter, Breathing, to criticism of the Dutch study. Timmons claimed that the Dutch team’s sensors react slowly to changes in blood levels of CO2, so it is difficult to conclude that hyperventilation does not precede an attack of symptoms. She also wrote that asking patients to undergo two HPVTs was ethically unacceptable, because of the test’s potentially unpleasant consequences. The Dutch study had, in fact, received ethical approval.

Timmons brought advanced copies of the newsletter to court on the day van Spiegel gave evidence, and sat in court as Nixon’s lawyers used it to challenge the Dutch research. Other European members of ISARP did not receive the newsletter until a week later. It was “an ambush”, says van Spiegel, “discrediting me and our team’s published research…It is not acceptable behaviour”.

Last month, Ley said he had not read the Dutch study when it was published, nor had he tried to publish his views in The Lancet, which would have been the conventional place to criticise the research. He declined to say whether he still supported the HVS theory: “I don’t believe newspapers are the proper place to discuss this.” Timmons has declined to talk to New Scientist.

Van Spiegel and others are concerned at the appearance that Ley and Timmons tried to harness ISARP’s reputation to Nixon’s cause. Kees Wientjes, chairman of the Dutch chapter of ISARP, has asked them to account for their conduct. To date, they have not replied to his questions and a clash seems inevitable. “A scientific society that gives—due to actions of its most prominent board members—the impression to foster and promote certain scientific beliefs and to discourage or even discredit research critical of such beliefs is doomed,” says Wientjes.

Schism

Wientjes says a “schism” has opened within ISARP between “a camp of exponents of the ‘classical’ belief in hyperventilation as a physiological problem underlying an endless range of medical and psychological disorders, and a camp of researchers primarily interested in an empirical and evidence-based approach”. The rival camps will meet next month to debate the definition of HVS. Following the Nixon affair, “it seems inevitable that the position of the ‘believers’ is weakened”, says Wientjes.

The story of Nixon and HVS raises questions about how to police the quality and ethics of research, and the standard of peer review in medical journals. For HVS, lessons need to be learnt quickly. If the clash between “believers” and empiricists continues, it bodes ill for medicine and patients. “A clear treatment strategy is vital,” says Gardner. Without it, a host of “otherwise fit and often young patients” will become or remain chronically ill.

Duncan Campbell is a journalist and broadcaster