Correspondence
Do Medical Journals Suppress Information?
N Engl J Med 1993; 328:510-511February 18, 1993
- Article
To the Editor:
In response to charges of intellectual suppression reported in Newsweek,1 your editorial (Oct. 22 issue)2 comments on a paper rejected by the Journal that was recently published elsewhere3. This work presents the hypothesis that neonatal exposure to protoporphyrin-activating fluorescent lighting is a contributing and preventable cause of childhood leukemia. Your commentary discloses a portion of the letter rejecting an earlier version of this paper four years ago, which noted that the hypothesized relation was “entirely speculative and, while interesting, would not be suitable for publication without substantial supporting data”2.
Because the usual confidentiality of peer review has been broken in this matter, readers should also be privy to other statements in the correspondence from the Journal. One reviewer stated, “This is an intriguing idea that can be readily tested.” And the Editor commented, “I am certain you will have no difficulty getting this provocative work published elsewhere.”
Your selective commentary takes an inappropriate swipe at both the journal that did publish the paper and the hypothesis, arguing that “the fact that the study was published several years later in an obscure journal if anything supports our editorial decision to reject the paper.” This is flat-out wrong on two counts. First of all, the eventual publication of our hypothesis succeeded because it managed to capture supporting data, generated in response to professional presentations of the concept. Our hypothesis combines solid epidemiologic data related to patterns of childhood leukemia with seemingly unrelated phenomena pertaining to the ontogeny of the hematopoietic system, photoactivation of protoporphyrins, and ensuing DNA damage. Second, the periodical in which our hypothesis recently appeared, Cancer Causes and Control,3 is not an obscure journal; under the Rapid Communications of Oxford Limited program, this new journal provides front-line research results on the causes and treatment of cancer, with world-class reviewing editors, many of whom also serve as reviewers for your journal.
Recent studies of childhood leukemia in Sweden and the Netherlands have provided some support for our hypothesis, finding that children with less exposure to fluorescent lighting had significantly lower rates of the disease4,5. Other tests of this hypothesis are under way, sponsored by Health and Welfare Canada and the Leukemia Fund in England. Only by further study will we learn whether protoporphyrin-activating lighting constitutes a completely preventable contributing cause of some proportion of childhood leukemia, a disease with an increasing incidence6 and poorly understood causation.
A Catch-22 underlies your commentary. How can clinical data be gathered to support new hypotheses if such hypotheses are not afforded some outlet for publication in journals with wide readership? Making room for innovative work that falls outside conventional paradigms presents a challenge to all journals. The Journal should provide for the occasional publication of medical hypotheses in order to stimulate advances in biomedical theories.
6 ReferencesDevra Lee Davis, Ph.D., M.P.H.
National Research Council Commission on Life Sciences, Washington, DC 20418Shmuel Ben-Sasson, Ph.D.
Hubert H. Humphrey Center for Experimental Medicine and Cancer Research, Hebrew University, Jerusalem, Israel1
Begley S. Is science censored? Newsweek. September 14, 1992:63.
2
Kassirer JP . Do medical journals suppress information? N Engl J Med 1992;327:1238-1238
Full Text | Web of Science | Medline3
Ben-Sasson SA ,Davis DL . Neonatal exposure to protoporphyrin-activating lighting as a contributing cause of childhood acute lymphocytic leukemia. Cancer Causes Control 1992;3:383-387
CrossRef | Web of Science | Medline4
Zack M ,Adami HO ,Ericson A . Maternal and perinatal risk factors for childhood leukemia. Cancer Res 1991;51:3696-3701
Web of Science | Medline5
van Steensel-Moll HA ,van Duijn CM ,Valkenburg HA ,van Zanen GE . Predominance of hospital deliveries among children with acute lymphocytic leukemia: speculations about neonatal exposure to fluorescent light. Cancer Causes Control 1992;3:389-390
CrossRef | Web of Science | Medline6
Annual cancer statistics review 1973-1989. Bethesda, Md.: National Institutes of Health, 1992.
To the Editor:
I disagree with one point in your editorial about the suppression of information by medical journals. You stated that the fact that a study was rejected by the Journal but was subsequently accepted by an obscure journal supports your editorial decision. Whereas science is (or should be) objective, the review process is quite often subjective. I have had a manuscript rejected by one journal because of a “poor data source,” only to have the same manuscript accepted elsewhere because of, among other things, an “excellent data source.”
Because it is so prestigious, the Journal can publish excellent studies; however, the rejection of a study by the Journal and its subsequent acceptance by an “obscure” one does not necessarily undermine the validity or importance of the study.
Joseph F. Antognini, M.D.
University of California, Davis, Davis, CA 95616Author/Editor Response
Editor's reply:
The confidentiality of peer review was broken not by the Journal, but by Dr. Ben-Sasson in his comments published in Newsweek. We have never taken a swipe at his hypothesis; we simply were not interested in publishing a hypothesis that was so meagerly supported by data. In fact, we applaud studies that test his hypothesis, and we will be interested in the results. The journal in which the paper was published, Cancer Causes and Control, is relatively new. It is only in its third edition and is not widely known. We certainly have no argument with journals such as Cancer Causes and Control that publish hypotheses and label them as such. We simply chose not to do so ourselves.
Dr. Antognini makes a cogent point. It probably would have been more accurate to say that the publication of a paper by a second journal provides little or no evidence about the validity of a previous decision to reject the paper. The interval of four years between rejection and publication in another journal, however, suggests that the paper required substantial reworking, that there were other rejections, or both.
Jerome P. Kassirer, M.D.
