We thank the various correspondents for their thoughts on our paper. When looking at cohort data is it crucial that periods at risk are used for the denominator when calculating rates, not the number of women in each group. The periods of observation for ever users in both the main and GP observation datasets were higher than those for never users, so even though the total number of deaths was higher in the ever user group the rate of death was lower than that among never users in the main dataset (and near unity in the GP observation dataset).

Professor Brind feels that common sense should have told us that when there are material differences in statistical trends we should use the smaller, better dataset. We presented both datasets so that readers can decide for themselves which one(s) they wish to use. In neither dataset was there substantial evidence of an increased risk of all cause mortality among ever users. Thus, even in the smaller GP observation dataset the upper 95% confidence interval for the relative risk of any death between ever and never users was just 1.10. We believe that the two datasets found different results because a higher proportion of the periods of observation in the ever user group in the GP observation dataset related to current and recent pill use, than in the main dataset. This explains why the GP observation dataset results found an increased risk of all circulatory disease and an increased risk of breast cancer in ever users who stopped oral contraception 5-9 years previously. Similarly, in the age-stratified analyses of the main dataset, ever users aged <30 years had much more current and recent usage than older ever users women. Thus, it was not surprising that ever users <30 had nearly three times the risk of death than similarly aged never users. We have previously shown that most of the adverse mortality effects of oral contraception occur in current and recent users, effects which diminish with time since stopping 1. Our latest findings are compatible with our previous publications, and broadly in line with the collective evidence from other studies. A substantial proportion of the periods of observation in the ever user group in the main dataset relates to pill usage many years in the past. Unlike Professor Brind, we feel that it is valuable to have the information provided by this dataset. We continue to interpret the results as not suggesting a substantial increased overall risk of death among ever users, especially since many events occurred long after the pill was stopped and so any medical treatment and subsequent death certification is unlikely to have been influenced by the doctor’s knowledge of a woman’s pill usage.

We have tried to be careful when interpreting our results, both in the BMJ publication and in subsequent discussions with the media. Thus, we have stated clearly that the reduced overall risk of death in ever users in the main dataset may be due to selection processes or residual confounding rather than a direct effect of oral contraception. Indeed in the paper we reminded readers of the striking difference between clinical trial and observational data in relation to hormone replacement therapy. We did not collect information of body mass index at recruitment and so failure to adjust for this variable might explain our results. Similarly, as Dr Grant correctly points out, we could not adjust our main dataset results for any hormone replacement therapy used by the women. We have previously shown that oral contraceptive users in our study were more likely to use hormone replacement therapy than never users2. As a consequence, any hormone replacement therapy-related adverse effects would tend to be stronger in the ever user than the never user group. Thus, failure to adjust for hormone replacement therapy use would tend to diminish rather than exaggerate differences between ever and never users of oral contraceptives.

We continue to be perplexed as to why our study found an increased risk of violent deaths among ever users. Dr Robert’s explanation is intriguing although such an effect would have to be strong and persisting to account for a continuing risk in our study.

References

1. Beral V, Hermon C, Kay C, Hannaford P, Darby S, Reeves G. Mortality associated with oral contraceptives: 25 year follow of cohort of 46 000 women from the Royal College of General Practitioners’ Oral Contraception Study. British Medical Journal 1999; 318: 96-100.

2. Moorhead T, Hannaford P, Warskyj M Prevalence and characteristics associated with use of hormone replacement therapy in Britain. British Journal of Obstetrics and Gynaecology 1997; 104: 290-297.

Competing interests: The Centre of Academic Primary Care has received payments from Schering Plough and Wyeth Pharmaceutical for lectures and advisory board work provide by PCH. None of the other authors have a conflict of interest to declare.

Unfortunately the partly drug funded Royal College of General Practitioner (RCGP) contraceptive pill report makes claims of reduced mortality which are flawed and difficult to entangle.1

Contraceptive pills cannot magically prevent deaths.

Data omissions include the following: the RCGP pill study was designed in 1968 to have 23 000 “ever” and “never” takers but deaths were flagged for 21 936 users and only in 13 168 never takers; by 1972 more pill takers had been lost than controls so more takers were added; some “never” takers were moved into the taker group; when deaths were flagged in 1977 most current use (average age 29 for 44 months) would have already stopped; information on fertility drugs or HRT (also progesterones and oestrogens) were not revealed for the whole duration of the study; some HRT use was known until 1996 but 3 out of 4 deaths occurred in the next 10 years. Although more deaths were in “never” pill takers over age 50, these women were significantly older when recruited and a larger proportion would have taken HRT which causes increases in mortality.

Deaths increased three times more in “ever” takers under age 30 than in young “never” takers. GP observed “ever” takers had significant increased mortality rates compared with “never” takers for all circulatory diseases, cerebrovascular disease, other circulatory diseases (thrombosis), and violence (perhaps reflecting previously increases in mental illness and marital break ups in “ever” takers). A much vaunted ovarian cancer reduction depended on 14 deaths in “ever” takers and 29 deaths in “never” takers and 75 deaths in each group in the full data set. Whether these women were taking fertility drugs or HRT, which can increase the risk of ovarian cancer, is unknown.2,3

The study Flow Chart shows that from 1996 to 2007 “ever” pill takers in the GP observation set deaths had 10.3 times more deaths compared with 6.5 times in “never” takers. Women who left GP observation younger than age 38 had 8.7 times more deaths if “ever” users but no increase in deaths occurred in “never” takers. “Ever” takers of oral contraceptives had 2864 deaths; 1312 from all cancers including 312 breast and 258 lung cancers; 763 from circulatory diseases; and 156 deaths from violence.

Breast cancer incidence and mortality increases and decreases over the last 50 years match similar changes in progestogen plus oestrogen use.4

For 50 years studies have shown adverse biological effects including blood vessel and enzyme changes. In addition numerous epidemiological studies, including the prestigious Women’s Health Initiative randomised study, have confirmed that taking progestogens and oestrogens increases risks of cancers, vascular and mental diseases.5 Previous RCGP publications found deaths increased from breast, cervical, lung and liver cancers and vascular diseases in takers. However, most studies underestimate risks because of drop outs, losses and confusion about the effects of taking hormones for different reasons at different ages.

Hormone use for any reason does cause major public health problems. Omissions in study data give false reassurance to young women who are particularly at risk.

1 Hannaford PC, Iverson L, MacFarlane TV, et al. Mortality among contraceptive pill users: cohort evidence from Royal College of General Practitioners’ oral contraception study. BMJ 2010:340:c927.

2 Grant ECG. Increased risk of serous ovarian cancer following clomifene fertility treatment.http://bmj.com/cgi/eletters/338/feb05_2/b249#208597, 11 Feb 2009

3 Zhou B, Sun Q, Cong R, Gu H, Tang N, Yang L, Wang B. Hormone replacement therapy and ovarian cancer risk: a meta-analysis. Gynecol Oncol. 2008;108:641-51.

4 Colditz GA. Decline in breast cancer incidence due to removal of promoter: combination estrogen plus progestin. Breast Cancer Res. 2007;9:108.

5 Chlebowski RT, Schwartz AG, Wakelee H, Anderson GL, Stefanick ML, Manson JE, Rodabough RJ, Chien JW, Wactawski-Wende J, Gass M, Kotchen JM, Johnson KC, O'Sullivan MJ, Ockene JK, Chen C, Hubbell FA; Women's Health Initiative Investigators. Oestrogen plus progestin and lung cancer in postmenopausal women (Women's Health Initiative trial): a post-hoc analysis of a randomised controlled trial. Lancet. 2009 Oct 10;374(9697):1243-51.

Competing interests: None declared

Table with data from RCGP Contraceptive Pill Flow chart. Flagged Deaths in 1996 and 2007

				Numbers of Deaths/Total women  (Increases)       
 		   1996  		                2007 
 	         Ever	         Never	     Ever	    Never
Under GP
observation 
in 1996	     62/6538	    73/3956	  640 (x10.3) 478 (x6.5)
Had left GP
observation
age <_38 _="_" _74="_74" _7761="_7761" _23="_23" _4154="_4154" _641x8.7="_641x8.7" _23x0="_23x0" pre="pre"/>

This very large and long term study on safety of oral contraception once again confirms long term safety and efficacy of oral contraceptives. The results are very promising that ever users of oral contraceptives had infact a lower mortality rate as comparted to never users. This should silence any critics once for all. The results also highlight that most cancers including gynaecological and bowel cancers were less common in oral pill users as compared to never users.This is especially relevant as there has been question mark over safety of estrogens on human health after their adverse effects were noted with estrogen replacement therapy in elderly women.In fact oral pills also increase bone mineral density and should reduce fracture rates amongst users as compared to never users.

Traditionally it is believed that oral pills may increase incidence of cervical cancer but the present study found its lower incidence in the users.However, the authors did not mention about breast cancer which many sceptics still feel may be increased with use of oral pills as most breast cancers are estrogen receptors positive and this cancer is the commonest cancer amongst women in the world.

The results of this large study are really welcome and promising to highlight long term safety of oral pills which are the most effective method of contraception and save many womens lives from complications of unwanted pregnancies. In fact not using them causes many unwanted pregnancies especially in developing countries with unsafe abortions causing preventable maternal mortality. There is a real need to remove the various false myths about oral pills in developing countries to increase their use in which direction this study is a huge step.

Competing interests: None declared

As I expected, the paper had International resonance and as I was reading am (Italian) GP daily news magazine, the message taken home has been: contraceptive use does not increase mortality. Besides to compliment Prof Hannford and colleagues for the "titanic" work always associated with such big cohort studies, I must remember that the paper does not strongly support that conclusion. Its very likely, in my opinion, that there are differences between the 2 groups other than the use of contraceptivw. We may argue that some patient not taking contraceptive may have done so as they were not engaged in relationship(s) because of health related issues. It may be assumed that some patients may have not taken contraception because of already known important risk factors or established medical problems.

Regards,

E Cervoni, MD

Competing interests: None declared