Tom Treasure Weblog

Reading the most enjoyable, varied and stimulating blogs from Dr Bob Replogle I realise that I have again taken a rather dry surgical subject. But seeing the calendar at November 2005 I wanted to catch the fiftieth anniversary to tell this story. To lighten the tone I will close with a poem by Hilaire Belloc. If you read nothing else, read that.

Fifty years ago this month, in November 1955, the thoracic surgeon Geoffrey Flavell performed a remarkable clinical experiment. He asked a patient with lung cancer if he would consent to a thoractomy during which Flavell would only divide the vagus; he proposed re-opening the chest incision a week later to perform the lobectomy (Lancet 1956; 270:260-262 ). In May this year I gave the inaugural Geoffrey Flavell Lecture at the Society of Apothecaries in London and that prompted me get out the original Lancet paper. The man had severe hypertrophic osteoarthropathy (HPOA) Flavell wanted to test the hypothesis that the vagus nerve was the afferent pathway of a mechanism that caused clubbing and the more severe changes of HPOA. The man had first developed painful swelling of the ankles in August and over the next couple of months the knees, then the wrists, elbows and fingers also became stiff and painful. “Radiography showed typical hypertrophic osteoarthropathy of all the long bones”. Remember that this was about 20 years before the first CT images and Flavell relied for assessment of operability on simple radiology and his rigid bronchoscope. There was a small opacity in the upper lobe of the left lung which was diagnosed as lung cancer on sputum cytology. “The patient agreed to a two-stage operation”.



Dr. Geoffrey Flavell

At the first operation on 23rd November 1955 Flavell disturbed nothing but the vagus nerve and divided it just distal to the recurrent laryngeal nerve. The following morning the patient was asked if he felt any different. “Goodness me, yes” Flavell quotes “all my pains are gone; and look – I can bend my knees for the first time in weeks!”. The swelling went in 48 hours and the pain never recurred. A week later Flavell removed what proved to be a squamous carcinoma without lymph node metastases.

The background as Flavell wrote was this: “It has long been known that resection of the growth is followed instantly by cessation of all pain …” This is accepted whenever I have discussed it with colleagues; it is a striking and consistent observation. Flavell made the further observation that pain is also relieved in patients in whom resection had failed, but a preparatory dissection around the hilum had been performed. He describes two such cases in this paper. In another two cases, both on the right, he divided the vagus above the azygos vein, the last through a minimal thoracotomy. In all cases, the patients’ pain was completely abolished. In his conclusion Flavell proposed division of the vagus nerve in inoperable cases with the objective of relieving pain.

As a resident at the London Chest Hospital in 1978, I was studying the teaching collection of x-rays in the radiology department. I chanced upon the films an inoperable right sided lung cancer and a pathology laboratory report of a confirmatory biopsy of the vagus nerve obtained through a mediastinoscope. It was impressed upon my memory. Reaching and dividing the vagus – and sending a bit to the lab – seemed to me not something that would have been attempted by the faint hearted or unskilled. The surgeon was a resident called Magdy Yacoub. It was this chance discovery of the story that prompted my interest in HPOA and my curiosity remains. The young Yacoub reported this and another case (Br J Dis Chest 1965; 59:28-31, Br J Dis Chest 1966; 60:144-147). One was a man who was admitted to hospital in 1962 with left sided cancer which had already destroyed the recurrent laryngeal nerve. The pain from HPOA was so severe that he could not sleep. Division of the vagus in the neck gave dramatic relief. In the second case in 1964 the patient had lost the use of his hands due to the severity of HPOA. This was the man in whom Magdy passed a mediastinoscope through the usual neck incision divided the vagus nerve in the tracheo-oesophageal groove. “The joint and bone pains subsided immediately, and five days later he could close his fists easily”.

Flavell and Yacoub reviewed the evidence for proposed mechanisms. On the basis of these cases and their review of the experimental and clinical evidence they believed that the ipsilateral vagus nerve was the key afferent pathway. While the observation that pain is consistently relieved by resection is generally agreed, standard clinical texts on lung cancer shy away from this vagal hypothesis, stating that the mechanism is unknown. The possibility of a vagal mechanism is mentioned in some texts but usually no source is cited. Little credence is given to the vagal contribution and various humoral mechanisms are postulated. Since 2000 there are fewer than 100 references to HPOA in PubMed, predominately case reports of rare associations and strange phenomena. Lung tumours cause a wide range of paraneoplastic syndromes due to the release of proven or postulated substances and a neural and humoral mechanisms could co-exist. However, the vagal mechanism is not favoured. It appears to be largely discounted. Why?

Is there a question about the reliability of these surgical reports? This seems to me unlikely. Cardiothoracic surgery was then and remains a very public specialty. The London Hospital and the London Chest Hospital were both major contributors to teaching, training and clinical reports in the 1950s and 1960s. Flavell named each referring physician in his Lancet paper. Yacoub acknowledged his senior colleague Jack Belcher when he wrote in the British Journal of Diseases of the Chest (a journal which one of my mentors in the 1970s used to refer to as the Brompton Hospital “school mag”). Erroneous or over enthusiastic assertions could hardly have gone unremarked.

There is the possibility that the dramatic relief reported by Flavell and Yacoub was a placebo effect - a result of suggestion. There are good reasons to consider that possibility. When we see lung cancer patients with HPOA at least one doctor, and probably several, will tell them “That pain will go away when we take the cancer out”. A thoracotomy would surely qualify as an intervention sufficient to prompt a placebo effect. And yet the effect is consistently reproducible and the physical and radiological changes regress, along with the pain. A placebo effect seems unlikely.

HPOA is generally regarded as at the severe end of the clinical spectrum of the physical changes that are seen under the general heading of finger clubbing. Clubbing is said to have been described by Hippocrates in his account of empyema and is seen in about a third of cases of lung cancer. The incidence of HPOA in lung cancer was about 3% (9 out of 280) in a consecutive series seen between about 1970 and 1975. The occurrence of “Hypertrophic pulmonary osteoarthropathy without clubbing of the digits” (Skeletal Radiol 2001; 30:652-655) is so rare as to be published only in as few as four case reports.

In many discussions of the subject which I have had with very well informed physicians, Flavell’s evidence is played down and it is because the mechanism seems to them to be implausible. The preferred theory is that there is some form of tumour derived growth factor which is responsible. Although the surgical “experiments” of Flavell and the clinical application by Yacoub provide strong empirical evidence for a central role for the vagus, the theory is not liked because it does not fit with any known physiological mechanism or that appears plausible in the light of current accepted scientific knowledge. There is a danger there. In the 1840s Semmelweiss provided evidence that hand washing with chlorinated water between the post mortem room and the labour ward dramatically reduced the death rate from puerperal sepsis but his contemporaries were able to discredit him because the microbes had “never yet been seen”. I would accept that any parallel between the discrediting of Semmelweis and the forgetting of Flavell would be far too strong but the more gentle message from Belloc is worth recalling: “Oh! Let us never never doubt what nobody is sure about”.

If anyone has more evidence about HPOA and the vagus I would be interested to know. With modern minimal access approaches, perhaps we should consider offering this form of palliation in inoperable lung cancer.

THE MICROBE
by: Hilaire Belloc (1870-1953)
HE MICROBE is so very small
You cannot make him out at all,
But many sanguine people hope
To see him through a microscope.
His jointed tongue that lies beneath
A hundred curious rows of teeth;
His seven tufted tails with lots
Of lovely pink and purple spots,
On each of which a pattern stands,
Composed of forty separate bands;
His eyebrows of a tender green;
All these have never yet been seen--
But Scientists, who ought to know,
Assure us that is must be so...
Oh! let us never, never doubt
What nobody is sure about!

"The Microbe" is reprinted from More Beasts for Worse Children. Hilaire Belloc. Duckworth, 1912.

Posted by Tom Treasure at 07:08 AM | Permalink

Over the past few years, this question has become more pressing, both for the profession and for me as a surgeon. However long and hard I look at the problem I come to the same conclusion – we need a proper controlled trial.

For the twenty years until 2001 did a very broad range of work as a cardiothoracic surgeon. My cardiac surgery, in addition to the full range of elective and emergency surgery for ischaemic and valvular heart disease, included an increasing tertiary referral practice in aortic surgery and a period of some years in transplantation. Consequently in thoracic surgery, to somewhat limit my range, I left oesophageal surgery and chest wall reconstruction, for example, to full time thoracic colleagues. As for Mesothelioma, it was rare when I started and I was content to see the few cases that came along go to the pure thoracic surgeons for an opinion. Now two things have changed. Mesothelioma has become much more common and, having been invited back to Guy’s Hospital, my alma mater, as a General Thoracic Surgeon, I have had to take a position on this challenging problem.

The problem: Mesothelioma now kills more people a year than either melanoma or cancer of the cervix. The death rate will rise for 10-15 years. About 60,000 new cases are projected in the UK over the next 45 years. The same is true in Europe. The epidemic may be declining in the USA but in many countries in Asia and South America asbestos is still uncontrolled and there is bound to be a huge burden of illness.

Is there any hope of cure? Extrapleural pneumonectomy (EPP) promulgated in the 1970s by the British surgeon Eric Butchart is the radical surgery on offer in which the parietal pleura, mediastinal pleura, pericardium and the diaphragm are taken en bloc with the lung inside the envelope. If surgery alone had a self evident role it would probably have been established by now – or would it?

An historical digression: Consider the history of mitral valvotomy for rheumatic mitral stenosis. In the 1940s the proposed operation of valvotomy was authoritatively declared to be not only dangerous to the point of recklessness but also misconceived since any physician knew that it was the myocardium that was the problem; the stenosed valve was a non-contributory epiphenomenon. Within a few years of Bailey, Harken and Brock’s operations in 1948 it was universally accepted. Surgery had been first proposed fifty years earlier in 1898. It took fifty years to become established (Treasure T, Hollman A. The surgery of mitral stenosis 1898-1948: why did it take 50 years to establish mitral valvotomy? Ann R Coll Surg Engl 1995; 77:145-151).

The case for EPP: The best survival has been reported by David Sugerbaker in a subset of 31 patients who had three prognostically significant variables: epithelioid cell type, clear resection margins, no positive extrapleural nodes (JTCVS 1999). Their median survival was 51 months. No reference population anywhere has survival figures with a median in excess of four years. On the strength of these results, Sugerbaker’s team offer patients who meet these criteria radical surgery preceded usually by chemotherapy and followed by radical radiotherapy. Not to do so would be to deny 50% the chance of cure. They add weight to their case for EPP (or weaken it, depending on one’s point of view) by adding an argument ad hominem, accusing the other side (and this is an issue where sides are taken) of therapeutic nihilism and surely nihilism in this and all cancers must be a thing of the past?

The case against EPP: To perform EPP is to base a lot on just 31 patients retrospectively selected. It may all be a consequence of retrospective exclusion of the patients with unfavourable features. Let me explain. If one establishes favourable features from a data base, and redefines the groups on the basis of significantly favourable and unfavourable features, there will be a difference in favour of the better group. That is a statistical inevitability, leading to a circular argument. That is why in developing a risk model the data are randomly divided into a first test set to devise the risk model and the remainder are used to test its reproducibility.

Is it likely that the results will be replicated prospectively? Let us look at that again in clinical terms. You have identified the favourable features for longer survival. You do not operate on those without favourable features and confine your efforts to those most likely to survive. You then look at survival for the selected group operated on. They will probably do less well than the artificially selected group in the retrospective analysis but they may well do better than historical controls. But the historical controls are dominated by the very sort of cases that you elected to not operate on. Again, it becomes circular and self-serving, say the doubters.

Like those in favour, those against EPP seek to bolster their arguments with emotive language, answering the charge of therapeutic nihilism with references to quality of life. They emphasise the inescapable fact that trimodality therapy will take six to nine months. Perhaps it is inhumane to use up the best remaining months of someone’s life this way.

The basis of equipoise: So this is the situation I found as I took up my tools as a born again general thoracic surgeon. Which policy to adopt? The disease was becoming more common and we were diagnosing cases ever more frequently, sometimes several a week. Which camp was I in? Either way I faced a personal problem of integrity. Could I, as a sceptic, deny active treatment without the evidence to back up that decision or, as an enthusiast put patients through extreme treatment on a judgement made outside the evidence? Whichever of these two opposing views is espoused, and however well intended, the evidence is simply not there to support either. Are we entitled to a free for all, each side engaging in highly charged rhetoric?

My colleague Jules Dussek invited me to a meeting of his European Thoracic Surgical Club in 2001 soon after I joined him at Guy’s. On the programme was the Italian surgeon Maggi who related his EPP experience. Also present was the epidemiologist Julian Peto who had correctly predicted the epidemic in the The Lancet in 1995. Peto saw a randomised trial as the only way to address the uncertainty. He even had an acronym ready for the trial - MARS (Mesothelioma and Radical Surgery). Independently of Peto, the English surgeon David Waller and the Irish oncologist Ken O’Byrne had proposed a trial, so we joined forces.

One compelling reason for a trial is that the numbers of cases in years to come will place a heavy burden on resources. If treatments are effective we must fight to have them at our disposal. But what if they are not? The medical philosophers Brody and Miller have laid down the gauntlet with this strong challenge:

“So long as all the physician had to offer the patient was his own time and advice and a few herbs and powders, both medicine and society could comfortably claim that the physician’s duty of fidelity was owed solely to the individual patient. When physicians can, with the stroke of their pens, literally bankrupt the community, the community may no longer be able to tolerate that view of the physician’s duty. (J Med Philos. 1998;23:384-410)

The physician’s perception of where duty lies can lead to some anomalous situations. Respiratory physicians are often the first point of contact with patients, putting them in a powerful position to effectively veto the trial in their area, whether on the grounds of science or ethics. At the other end of the control chain is the surgeon. When presented with information about the trial some surgeons say it is “unethical” to offer this operation, others claim it is “unethical” to deny it to the patients who in their judgement might benefit. Other than quibbling with their casual use of the strong charge that colleagues are “unethical” I cannot argue convincingly against either.

The trial design: A trial which looks like “a lot” versus “a little” is hard for doctors and patients, hence the attempt to balance the arms. In the MARS trial EPP is sandwiched between chemotherapy and radiotherapy as that is how the best survival rates have been achieved. In MARS, the control patients are also given full active trimodality therapy; every treatment is available to the patient - short of EPP.

Is it ethical? The commonest objection we receive is on this basis. Yes, it is ethical. This trial has been extensively discussed in National and International meetings on many occasions during its development phase. It has passed through all the hoops to gain ethical acceptance (MREC) as well as scientific approval (CTAAC and NCRI) and funding (Cancer Research UK). Its ethics have been carefully considered.

Consider the alternative. What can be “ethical” in deliberately choosing to remain in uncertainty and thus consistently use treatments that fail to help - or are actively harmful? It is beyond any individuals clinical judgement to unravel cause and effect. There is a widely variable course in this disease, only the best cases are selected for radical treatment, the treatment has three components, all severe, and not all patients complete all three. Any multivariate analysis is beset with confounding. Whichever way I look at it, I come to the same answer – we need controlled trials. MARS has started as a small scale feasibility study. It is the first proper trial of EPP and it should not be the last.

Posted by Tom Treasure at 07:03 AM | Permalink