Organic: What's in a name?

The organic foods industry is booming: by one estimate, the market for organic foods is worth $4 billion annually and is expected to grow at a rate of more than 24% per year. Faced with the threat of pesticide exposures and other food safety problems, many consumers are turning to organic foods in hopes of finding a healthy alternative, but there is currently no consistency in organic food labeling and no guarantee that foods labeled as organic are actually grown and processed in a purely organic fashion. There is also controversy about whether the label "organic" covers such new technologies as irradiation and genetic engineering. As part of the 1990 Farm Bill, the U.S. Department of Agriculture (USDA) is working to develop a proposed rule on organic foods. The rule would regulate the allowable methods, practices, and substances used in producing and handling crops and their processed products. The first draft of the proposed rule, released in December 1997, met with unprecedented opposition, which centered around the fact that the proposal appeared to virtually ignore the recommendations of a standards board formed to assist in the rule's development. Other criticism opposed three practices put forward for comment by the USDA: irradiation, genetic engineering, and the use of sewage sludge in farming. Due to the vehemence of the opposition to its original proposal, the USDA has decided to rewrite the proposed rule. In preparation for that proposal, the USDA Agricultural Marketing Service released three issue papers in October 1998 for public comment. The 10,000-plus comments received in response to those papers will be incorporated into the second draft proposal, due out later this year.

during the period 1960-71 are presented and compared with an earlier series treated in the period 1947 -50. It is estimated that about 29°, of the patients treated in the later period are cured, in the sense that they have a life expectancy similar to that of the normal population, compared with just under 20% in the earlier period, but this improvement is mainly due to an increased proportion of Stage I cases in the later period. The percentage cured is discussed in relation to the ratio of deaths to registrations in the East Anglian Region and it is suggested that under-registration of deaths from cancer of the breast may occur.
AT THE END of the 1950s Dr Diana Brinkley and I carried out a study on the results of treatment of 704 patients with female breast cancer seen in the Cambridgeshire Area from 1947 to 1950 (Brinkley & Haybittle, 1959). The longterm follow-up of this series (Brinkley & 1-laybittle, 1975) suggested that just under 20% of all patients were cured in the sense that they had a life expectancy similar to that of the normal population. Nevertheless, although the death rate in the series between 20 and 25 years was almost the same as that of a normal population, 8/23 deaths after 20 years were from cancer of the breast, which is about 16 x the number that would be expected from canicer of the breast in the normal population. No cancer registry was operating in the Cambridge area at the time when the patients were treated and, although we endeavoured to collect as representative a group as possible of breast-cancer patients, we were aware that the series was probably deficient in operable cases treated by surgery only, and also perhaps in late Stage IV cases who were never referred anywhere for treatment.
If one looks at the natioinal figures for deaths and registrations of female cancer of the breast, the former are about 60% of the latter, which seems surprising if only 200o or less of breast cancer patients are cured. About one-tenth of the uncured patients might still be expected to die of some other disease before recurrence manifested itself but this would still lead, in a steady state situation, to deaths from cancer of the breast being about 72% of registrations. Also, the fact that many patients cured in terms of life expectancy still die of cancer of the breast would tend to increase this figure.
Since 1960 a Cancer Registry Bureau has been operating in the East Anglian Region, and it has also been possible to obtain from the G.R.O. the number of deaths from cancer of the breast in East Anglia. tion in this area, the dotted line in Fig. I shows there has also been a real increase in incidence per 100,000 females. Obviously the registrations in any one year give rise to deaths in later years, so that to get an idea of the success rate from these figures, it is more sensible to compare the deaths with the registrations at an earlier period. This is done in Table I   TABLE I where it can be seen that when deaths are compared with registrations 5 years earlier the ratio is about 60% ,, and if compared with registrations 10 years earlier about 70%1. But the median survival of the "uncured" patients is only about 3 years, so that deaths in any one year must be occurring in patients registered on average much less than 10 years earlier, and the 5-year ratio is probably the most appropriate. Thus, we still have for East Anglia at least a 1000 discrepancy in deaths/registrations if our "cured" group is only 200/. To investigate this discrepancy, a study has been made of results of treatment of all female breastcancer patients registered in the Cambridge Cancer Registry Bureau from 1960 to 1971 inclusive.

Patients treated in 1960-71
The first part of the investigation was to see whether there had been any marked changes either in clinical material or results over this 12-year period, and for this purpose the period was divided into 3 4-vear intervals. Table II shows the total number of patients and their mean age in each interval. The numbers reflect the increasing incidence already mentioned, and there was no obvious trend or significant change in the age of the patients during the period. Fig. 2 shows the numbers of patients in different clinical stages. The staging procedure used was uniform throughout the period and based on the TNM classification (IUCC, 1968 (Mantel, 1966), similarly for the earlier cases, Stages I and II only, Stag-es I anii( 1IIIhlve tenide(1 to inicrease, -whilst the niumbers in Stage 11 have decrease(1. Fig. 3 sho-ws this variation as a percentage of the total patients. The band of Stage II's has tended to narrow, -whilst the bands of Stages I andI IIl have tenIde(I to increase slightly, but on the -whole there hlas beeni no miarked chanlge in the dlistribution of stages over the 12-year period. Fig. 4 showVs the survival ctur'ves for each of' the 3 perio(ds. The curve f'or  (Coompar-isoii with 947-50) patienits On thie basis of this analysis of trhe individual 4-year periods, it seems reasonable to treat the whole group of patients fron 1-960 to 1971 as a single group for comparison with the 1947-50 series, and the survival cuLrves for the 2 periods are shown in Fig. 6. It is evi(lent that the more recently treated patients aire doing better, and the lifference between the 2 cuirves is highly significa,nt (x2 12 742; <0-000;5). \Vheui, however, the gtroUps being between the Stage II curve The main change in treatmer that occurred between the 2 per an increased use of simple rat] radical mastectomy as the prima ment for operable cases. It could seen in Stage II cases, but the results of a clinical trial started in 1958 (Brinkley & Haybittle, 1971) would not support this view. It seems more likely that the apparently improved result in the Stage II cases is an artefact caused by the difference in the staging procedures used. In the 1947-50 series Stage II was defined as a tumour confined to the breast without involvement of skin or pectoral muscle, but with palpable mobile nodes in the (704) axilla of the same size. No notice was taken of the tumour size. The TNM staging used for the 1960-71 series places breast tumours with diameters greater than 5 cm For 1947-50 automatically in Stage III. This would tend to improve the results in Stage II and also in Sta'e III, a trend which is annarent in Fig. 7.
Another change in the staging between the 2 series is in the effect of supraclavicular nodes. In the older series their presence placed a patient in Stage IV, whilst in the 1960-71 series a supraclavicular node placed a patient in Stage III. This would tend to improve results in Stage IV in the earlier series and this also is apparent in Fig. 7.
The cured group of 1960-71 patients A comparison of the stage distribution in the 2 series is shown in Fig. 8, where the most apparent difference is the higher number of unstaged cases in the earlier series. The majority of these were operated on, but the clinical findings before mastres for ectomy were inadequately recorded for when staging purposes. Their survival rates were very similar to those of the Stage II cases in the same series (Brinkley & by stage Haybittle, 1959). ppears in If we make the comparison with the unignificant staged cases excluded (Fig. 9) we can see '>0 005) that the later series has a higher propor-3s.
tion of Stage I cases and a lower proporat policy tion of Stages II and IV. The differences riods was in II's and IV's must to sonme extent be her than accounted for by the changes in staging ry treat-already mentioned, but the differences in be postu-I's is probably a genuine effect due to the known lack of comprehensive collection of early operable cases for the 1947-50 series, and also perhaps a real trend towards earlier referral. One must conclude that the overall improvement in survival results shown in Fig. 6 is probably a reflection of the different clinical composition of the 2 groups rather than of improved methods of treatment. Fig. 10 shows the survival curve for the 1960-71 series in relation to the expected survival of the normal population of the same age distribution. The follow-up has been too short to reach the situation where the two curves might be running parallel, although one can begin to see a suggestion of this. An estimate of the cured group by a mathematical model (Haybittle, 1965) gives a value of 29o%, and it does not seem unreasonable to suppose that the survival curve in Fig. 8 (Brinkley & Haybittle, 1968) compared with the 18°/ figure estimated by the model after a much longer follow-up (Brinkley & Haybittle, 1975).

T)ISCUSSION
The estimate of 29% cured in the 1960-71 series goes a long way towards explaining the anomaly of the deaths/registrations ratio mentioned at the beginning of this paper. If one-tenth of the uncured group die from other causes this would leave at least 63°/ of the total group dying from cancer of the breast, although this figure would be increased by any excess deaths from cancer of the breast in the "cured" group. The ratio of deaths to registrations 5 years previously was about 61% (Table   I) which is still lower than the percentage predicted from the cured group analysis, but perhaps not unreasonable in view of the inaccuracies of death certification. It does, however, require that such inaccuracies, if they exist, should lead to under-registration of deaths from cancer of the breast, rather than to over-registration.
I am very much indebted to Dr E. AM. Kingsley-Pillers, Director of the Cambridge Cancer Registration Bureau, for allowing me access to the data for the 1960-71 period, and to Dr Diana Brinkley for first drawing my attention to the ratio of (leaths to registrations in the national figures.