Correspondence
Preventive Health Services
N Engl J Med 1994; 331:1156-1158October 27, 1994
- Article
To the Editor:
Sox's article questioning the value of screening and treatment for localized prostate cancer (June 2 issue)1 provides a limited view. He incorrectly states that studies have not measured the accuracy of screening with the prostate-specific-antigen test in men without signs or symptoms of prostate disease. Such screening has been extensively studied in the target population of men who would participate in screening programs,2,3 and in this population symptoms did not correlate with an increased risk of prostate cancer.3
Sox also incorrectly states that the accuracy of prostate-specific-antigen screening is unknown because biopsies were not performed in asymptomatic men with prostate-specific-antigen concentrations below 4 ng per milliliter and negative findings on rectal examination. In fact, the rate of positive biopsy findings in such patients has been found to be 0 to 7 percent.4,5
Screening based on measurements of prostate-specific antigen increases the percentage of prostate cancers detected by 78 percent and doubles the percentage that are confined to the organ and thus curable.2,3 About 90 percent of these cancers have histologic features of cancers that would be expected to progress during the patient's life span.2,3
Sox questions the value of treatment, but one of the two supporting studies he cites involved a carefully selected, nonrepresentative patient population. The only valid conclusion of this study is that expectant management is an option for older patients with low-stage, low-grade tumors and a life expectancy of less than 10 years. The safety of watchful waiting has never been demonstrated in young patients. The other supporting study cited by Sox has been questioned because of the use of unrealistically low rates of metastasis in the statistical model. More realistic rates yield a two- to three-year quality-adjusted survival advantage for a 65-year-old man when radical prostatectomy is chosen over watchful waiting.6
Before the widespread use of prostate-specific-antigen testing, prostate cancer was usually diagnosed only in its incurable stages. Prostate-specific-antigen testing has reversed this trend. The results of treatment are favorable for patients with an early stage of disease and unfavorable for those with advanced disease. The only practical means of reducing morbidity and mortality rates is to detect and treat the disease early. In men with a life expectancy of 10 years or more, there is reason to believe that prostate-specific-antigen screening will save lives.
6 ReferencesWilliam J. Catalona, M.D.
Washington University School of Medicine, St. Louis, MO 631101
Sox HC Jr . Preventive health services in adults. N Engl J Med 1994;330:1589-1595
Full Text | Web of Science | Medline2
Catalona WJ ,Smith DS ,Ratliff TL ,Basler JW . Detection of organ-confined prostate cancer is increased through prostate-specific antigen-based screening. JAMA 1993;270:948-954
CrossRef | Web of Science | Medline3
Catalona WJ ,Richie JP ,Ahmann FR , et al. Comparison of digital rectal examination and serum prostate specific antigen in the early detection of prostate cancer: results of a multicenter clinical trial of 6,630 men. J Urol 1994;151:1283-1290
Web of Science | Medline4
Vallancien G ,Prapotnich D ,Veillon B ,Brisset JM ,Andre-Bougaran J . Systematic prostatic biopsies in 100 men with no suspicion of cancer on digital rectal examination. J Urol 1991;146:1308-1312
Web of Science | Medline5
Colberg JW ,Smith DS ,Catalona WJ . Prevalence and pathological extent of prostate cancer in men with prostate specific antigen levels of 2.9 to 4.0 ng./ml. J Urol 1993;149:507-509
Web of Science | Medline6
Beck JR ,Kattan MW ,Miles B . A critique of the decision analyses for clinically localized prostate cancer. J Urol 1994;152:Suppl:1894-1899
Web of Science | Medline
To the Editor:
In Table 1, Sox lists the recommendations concerning digital rectal examination for prostate cancer but does not include recommendations regarding digital rectal examination for colorectal cancer. Although it is clear that digital rectal examination is of limited value as a screening test for colorectal cancer,1 it is still part of standard practice. Is screening for prostate cancer the only indication for a rectal examination? What about in women?
Table 4 suggests that the only vaccines indicated in adults are pneumococcal, influenza, and hepatitis B vaccines; there is no mention of tetanus-diphtheria toxoid, rubella, or measles vaccines. Three expert panels (those of the American College of Physicians, the Canadian Task Force on the Periodic Health Examination, and the U.S. Preventive Services Task Force) have recommended the use of all these vaccines in adults (except for the measles vaccine, which the Canadian Task Force has omitted).2 We hope that readers will not interpret the omission of these interventions as a recommendation not to use them.
2 ReferencesBruce D. Bialor, M.D.
Tom J. Wachtel, M.D.
Mark J. Fagan, M.D.
Rhode Island Hospital, Providence, RI 029031
Preventive Services Task Force. Guide to clinical preventive services: an assessment of the effectiveness of 169 interventions. Baltimore: Williams & Wilkins, 1989:47.
2
Hayward RSA ,Steinberg EP ,Ford DE ,Roizen MF ,Roach KW . Preventive care guidelines: 1991. Ann Intern Med 1991;114:758-783
Web of Science | Medline
To the Editor:
Sox is appropriately cautious, but he does not emphasize that all the studies supporting screening for breast and colon cancer have used the surrogate end point of disease-specific survival rather than the preferable end point of overall survival.
In the Swedish overview of mammographic screening that Sox cites, Nystrom et al. note that, in spite of over 2.5 million patient-years of follow-up, the relative risk of overall mortality in the screened groups was 1.00 (95 percent confidence limits not given) “as expected from the relative importance of breast cancer as a cause of death... and the observed effect of screening.”1 The 0.12 percent decrease in mortality from breast cancer at 12 years will perforce translate into a somewhat lower decrease in overall mortality because of mortality from other diseases. The data of Mandel et al.2 on screening for colon cancer show a 0.30 percent decrease in mortality from colon cancer at 13 years in the group screened annually (95 percent confidence interval, 0.12 to 0.48 percent) but a 0.00 percent difference in overall mortality (95 percent confidence interval, 20.09 to 0.09 percent).
It is therefore evident that an individual patient has less than a 1-in-1000 chance of improved survival after 12 to 13 years of screening for breast or colon cancer. Given the small size of this effect, perhaps instead of the blanket recommendations made by expert panels, it would be more appropriate for physicians to discuss the data on the effectiveness of cancer screening with their patients and help them decide whether screening is appropriate for them.
2 ReferencesCarl D. Atkins, M.D.
242 Merrick Rd., Rockville Centre, NY 115701
Nystrom L ,Rutqvist LE ,Wall S , et al. Breast cancer screening with mammography: overview of Swedish randomised trials. Lancet 1993;341:973-978
CrossRef | Web of Science | Medline2
Mandel JS ,Bond JH ,Church TR , et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med 1993;328:1365-1371[Erratum, N Engl J Med 1993;329:672.]
Full Text | Web of Science | Medline
To the Editor:
A strong case can be made for screening for hereditary hemochromatosis, at least among white men. Hereditary hemochromatosis is a relatively common disorder with a prevalence among whites on the order of 5 cases per 1000.1 Screening based on the serum transferrin saturation and serum ferritin level is uncomplicated and has a high positive predictive accuracy,2 although the requirement of a liver biopsy to confirm the diagnosis is a cause for some concern. Phlebotomy is a simple and relatively inexpensive method for treatment. If started early enough, it can prevent irreversible complications, including cirrhosis, hepatic carcinoma, congestive heart failure, diabetes, and arthritis.
Screening has a favorable cost-effectiveness ratio and saves both money and lives, according to estimates based on a wide range of assumptions.3 Questions remain about the appropriate age at which men and women should be screened, the frequency of screening, and the cutoff point for the screening tests. Nonetheless, the evidence in support of screening for hereditary hemochromatosis is comparable to that for many diseases for which we routinely screen.
3 ReferencesJoseph Cappuccio, M.D.
Pradyumna Phatak, M.D.
Rochester General Hospital, Rochester, NY 146211
Edwards CQ ,Griffen LM ,Goldgar D ,Drummond C ,Skolnick MH ,Kushner JP . Prevalence of hemochromatosis among 11,065 presumably healthy blood donors. N Engl J Med 1988;318:1355-1362
Full Text | Web of Science | Medline2
Edwards CQ ,Kushner JP . Screening for hemochromatosis. N Engl J Med 1993;328:1616-1620
Full Text | Web of Science | Medline3
Phatak PD ,Guzman G ,Woll JE ,Robeson A ,Phelps CE . Cost-effectiveness of screening for hereditary hemochromatosis. Arch Intern Med 1994;154:769-776
CrossRef | Web of Science | Medline
Author/Editor Response
Dr. Sox replies:
To the Editor: Dr. Catalona claims that I misstated the case in saying that studies have not measured the accuracy of prostate-specific-antigen screening in the population that would be the target of screening: men without any signs or symptoms of urologic disease. To measure the accuracy of a test, one must perform that test (measurement of prostate-specific antigen) and a gold-standard test (prostate biopsy) in all study patients. In three of the four articles cited by Dr. Catalona, the population that underwent prostate biopsy was a selected subgroup of men who had undergone prostate-specific-antigen testing (those with borderline levels of prostate-specific antigen in one study, levels above 4 ng per milliliter in two studies, and a prostate examination suggestive of prostate cancer in all three studies). One cannot calculate a test's sensitivity and specificity for all patients if one does not perform the gold-standard test in all patients. In the fourth study1 the investigators performed biopsies in all the patients but stated (correctly) that the results should not be considered representative of routine screening for prostate cancer, since the men who were studied had sought referrals to urologists. I believe that the articles Dr. Catalona cites simply buttress my argument that the studies to date have not measured the accuracy of prostate-specific-antigen screening in asymptomatic men.
Dr. Bialor and his colleagues should recognize that neither the Canadian Task Force nor the U.S. Preventive Services Task Force recommends a digital rectal examination as a screening test for prostate cancer. They cite the lack of evidence that treatment of asymptomatic prostate cancer reduces the probability of death from prostate cancer. Dr. Bialor and his colleagues correctly call attention to my inadvertent failure to include the recommendation by all three expert panels that adults receive tetanus-diphtheria toxoid every 10 years.
Drs. Cappuccio and Phatak make the case for screening for hemochromatosis. It will be interesting to see whether any of the expert panels on preventive services take up their challenge to evaluate this practice.
Dr. Atkins raises an important issue: the effect of cancer screening on total mortality. Two of the best studies of cancer screening2,3 show a reduction in risk of approximately 30 percent for mortality from breast cancer and colorectal cancer but no effect on total mortality during a decade of follow-up observations. In recommending screening, expert panels imply that deaths from other causes are preferable to deaths from these two cancers. To prove this point, one would have to show that people preferred death from causes other than breast and colorectal cancer.
3 ReferencesHarold C. Sox, M.D.
Dartmouth-Hitchcock Medical Center, Lebanon, NH 037561
Colberg JW ,Smith DS ,Catalona WJ . Prevalence and pathological extent of prostate cancer in men with prostate specific antigen levels of 2.9 to 4.0 ng./ml. J Urol 1993;149:507-509
Web of Science | Medline2
Nystrom L ,Rutqvist LE ,Wall S , et al. Breast cancer screening with mammography: overview of Swedish randomised trials. Lancet 1993;341:973-978
CrossRef | Web of Science | Medline3
Mandel JS ,Bond JH ,Church TR , et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med 1993;328:1365-1371[Erratum, N Engl J Med 1993;329:672.]
Full Text | Web of Science | Medline
