Folic Acid Fortification and Colorectal Cancer Risk NaNa Keum, MS, Edward L. Giovannucci, MD, ScD Abstract: The U.S. has been reported as the only country experiencing a decline in incidence rates of colorectal cancer (CRC), despite increasing prevalence of CRC major risk factors, including the Western dietary pattern and obesity. This paper presents a hypothesis that improved folate status in the U.S. is the factor that could most likely explain the seemingly contradictory phenomenon, although a momentary increase in CRC incidence rates was observed in the later 1990s with the initiation of nationwide folic acid fortification. To corroborate this hypothesis, time trends in CRC incidence rates and death rates in the U.S. were plotted by age, race, and gender based on data from the Surveillance, Epidemiology, and End Results (SEER); data were analyzed by simultaneously addressing the following four critical factors: (1) a long induction period between improved folate status and potential protection of CRC; (2) a change in the U.S. Food and Drug Administration regulation in 1973 on the dose of folic acid allowed in supplements; (3) differential impacts of 1973 regulatory change and 1990s mandatory fortification by race; and (4) changes in CRC screening over time in the U.S. Although this type of analysis precludes a definitive conclusion, available evidence suggests that the increase in CRC incidence rates in the later 1990s is unlikely due to folic acid fortification and, assuming a time lag of a decade or longer to see a benefit on CRC, folate appears to be one of the most promising factors that could explain the downward trend of CRC incidence rates in the U.S. (Am J Prev Med 2014;46(3S1):S65–S72) & 2014 American Journal of Preventive Medicine
Introduction
C
olorectal cancer (CRC), as the third most commonly diagnosed cancer and the fourth-leading cause of cancer death worldwide,1 affects predominantly developed countries with the Western lifestyle.2 Although incidence rates of CRC have generally increased in developing countries and have stabilized in developed countries in recent decades, notably, the U.S. is the only country experiencing a decline in CRC incidence rates.3 This decline, if not all due to screening, indicates the presence of increasing prevalence of protective factors specific to the U.S. One study that compared CRC incidence rates observed with those predicted by MISCAN-COLON model, a simulation model that estimated secular trends of CRC if changes in the prevalence of risk factors or screening practice had not From the Department of Nutrition and Epidemiology (Keum, Giovannucci), Harvard School of Public Health; and the Channing Division of Network Medicine (Giovannucci), Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts Address correspondence to: Edward L. Giovannucci, MD, ScD, Department of Nutrition and Epidemiology, Harvard School of Public Health, 665 Huntington Avenue, Boston MA 02115. E-mail: [email protected]. edu. 0749-3797/$36.00 http://dx.doi.org/10.1016/j.amepre.2013.10.025
& 2014 American Journal of Preventive Medicine
occurred, showed that about half the decline was accounted for by changes in risk factors and the rest by screening.4 One potential factor is folate (vitamin B9 found naturally in foods), which is suggested to protect against CRC by preventing aberrations in DNA synthesis (DNA uracil misincorporation) and irregularities in DNA methylation (DNA hypomethylation) that promote colorectal carcinogenesis.5 Folate status has increased substantially in the U.S. resulting from a change in U.S. Food and Drug Administration (FDA)’s regulation in 1973 on the dose of folic acid (the synthetic form of folate) allowed in supplements6 and from the fortification (140 μg folic acid/100 g) of flour and grains with folic acid in the 1990s.7 Of note, although the fortification became mandatory in 1998, it virtually started from 1996 because of voluntary fortification by food industries immediately after the release of the new regulation in 1996.7 Past analyses that overlaid the secular trend of CRC with changes in national folate intake over time created a concern that excessive folate intake might promote CRC rather than preventing it.7,8 However, these analyses7,8 are limited by their failure to simultaneously address the following four critical factors: (1) a long induction period between improved folate status and potential protection of CRC; (2) a change in FDA’s regulation in 1973 on the
Published by Elsevier Inc.
Am J Prev Med 2014;46(3S1):S65–S72 S65
S66
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
dose of folic acid allowed in supplements; (3) differential impacts of 1973 regulatory change and 1990s mandatory fortification by race; and (4) changes in CRC screening over time in the U.S. Thus, this article will examine the CRC trend in the U.S. in relation to folate intake by simultaneously accounting for the aforementioned four factors and consider the potential benefits as well as risks of increased folate status on CRC.
Long Induction Period A long induction period deserves special attention in the study of CRC because the progression from normal cells of the colon and rectum, to small adenomas, to large adenomas, and finally to adenocarcinomas, occurs over a long duration, up to 30–40 years.5 The presence of a long induction period between adequate folate status and reduced risk of CRC is evident from both molecular mechanistic9 and epidemiologic studies.10,11 In a study that measured total folate concentration in colonocytes of adenoma patients (cases) and individuals without adenomas (controls), there was a significant decrease in colonocyte folate levels from the normal tissue of controls, to the normal tissue of cases adjacent to adenomas, and to the adenoma itself.9 The lower folate levels corresponded to increases in uracil misincorporation and DNA hypomethylation even in normal-appearing, non-neoplastic tissue of cases.9 As it takes at least 10 years in general to progress from non-neoplastic tissue to cancer diagnosis,12 the presence of such genetic alterations in non-neoplastic cells in the proximal field of adenomas implies that macroscopically undetectable abnormalities involving folate deficiency occur at least a decade before CRC is apparent. Epidemiologic evidence also supports a substantial time lag of at least 12–16 years between folate intake and CRC risk.10,11 Such data suggest that a protective effect of increasing folate intake would be reflected in incidence rates of CRC only after allowing for the estimated minimum induction period of 12–16 years.
Change in FDA Regulation on Dose of Folic Acid Although focus has been typically on mandatory folic acid fortification in the 1990s, a marked increase in folic acid intake dates back to 1973 when the FDA increased the daily maximum dose of folic acid allowed in supplements and foods from 100 to 400 μg6 (a recommended dose for women of childbearing age to prevent birth defects).13 Although its impact was not widespread, approximately one quarter of the U.S population using vitamin supplements in 1973 experienced a significant increase in folic acid intake10 along with an additional rise from fortified breakfast cereals. Yet, no previous
study to our knowledge has considered how the increase in folic acid consumption in 1973 relates to population incidence and death rates of CRC. More complete pictures would be obtained if the secular trend of CRC is examined in conjunction with both FDA’s regulatory change in 1973 and mandatory fortification in the 1990s while accounting for a long induction period.
Differential Impacts of 1973 Regulatory Change and 1990's Mandatory Fortification by Race In consequence of the 1973 change, the major group of individuals whose folate status was expected to improve was supplement users. According to the National Health and Nutrition Examination Survey I (1971–1975), the prevalence of supplement use among the adults aged 20– 74 years was 27.5% in men and 37.9% in women, and whites were the major users.14 Thus, any benefit associated with an improved folate status would be more pronounced among whites. Although the fortification policy in the 1990s led to a population-level increase in total folate intake by approximately 100 μg/day, with the largest increase for whites,15 pre-fortification daily total folate intake of adults aged 45 to 64 years was lowest in blacks (242 μg/day in men and 206 μg/day in women).15 Given considerable evidence that additional folate intake results in the greatest reduction in CRC risk among individuals at low folate status at baseline,10,16,17 blacks had the greatest potential to benefit from the fortification. In summary, variations by race in the improvement of folate status predict that the impact of respective folic acid policy on CRC risk would differ by race as well. This emphasizes the critical need for investigating the temporal association between folate intake and CRC by race.
Changes in CRC Screening Over Time in the U.S. Understanding how screening influences incidence rates of CRC over time is crucial to teasing out the effect of screening from that of folate intake. A landmark trial18 that evaluated the effect of sigmoidoscopy showed that approximately for the first 3 years since the baseline randomization, incidence rates were higher in the screening group than in the usual care group because of an increased detection of asymptomatic CRC by screening. However, after Year 3, incidence rates became and remained lower in the screening group presumably because the removal of precancerous adenomas detected by screening led to the primary prevention of CRC. This competing reduction of CRC incidence offset and then outweighed the initial rise due to an increased detection. Taken together, it can be inferred that the initial increase in screening rates is immediately followed www.ajpmonline.org
S67
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
70
Rate per 100,000
60 50
Rate per 100,000
350 300 250 200 150 100 50 19 7 19 5 77 19 7 19 9 8 19 1 83 19 8 19 5 87 19 8 19 9 91 19 9 19 3 95 19 9 19 7 99 20 0 20 1 03 20 0 20 5 07 20 09
0
Year of diagnosis
B. Blacks
450 400 350 300 250 200 150 100 50 0
7 19 5 7 19 7 79 19 8 19 1 8 19 3 85 19 8 19 7 8 19 9 9 19 1 93 19 9 19 5 9 19 7 99 20 0 20 1 0 20 3 05 20 0 20 7 09
Based on data from the SEER, the overall secular trend of CRC in the U.S. was plotted. The trend was further plotted by age, race, and gender. Figure 1 presents the overall secular trend, which shows a sudden increase in CRC incidence rates in the mid-1990s. This is the figure that raised the concern that excess folate intake due to the nationwide fortification might promote colorectal carcinogenesis.7,8 However, Figures 2–4 show substantial heterogeneity in trends by subgroups of age, race, and gender, which underscores that interpretation of the temporal relationship deserves more delicate explanations.
400
19
Procedures
A. Whites
450
Rate per 100,000
by a momentary increase in incidence rates without a time lag for the first few years. Of note, the initial rise in incidence rates in the screening group did not translate into higher death rates. Mortality, as a function of both incidence and prognosis, can be reduced by decreasing incidence or improving prognosis. Screening reduces mortality through both mechanisms: the detection and removal of precancerous lesions lead to the primary prevention of CRC and, thus, lowers incidence; an early detection of asymptomatic CRC improves prognosis. Hence, the decreasing trends in mortality in spite of a momentary rise in incidence rates, which we termed as the screening bump, serve as the characteristic feature that distinguishes the effect of screening from the effects of risk factors that elevate both incidence rates and death rates in most cases.
Year of diagnosis Age (years):
20–49
50–64
≥65
Figure 2. Age-specific incidence rates of CRC for both genders by race (1975–2009): (A) whites, (B) blacks Note: Cancer sites include invasive cases only unless otherwise noted. Rates are per 100,000. Regression lines are calculated using the Joinpoint Regression Program Version 3.5, April 2011, National Cancer Institute. Data source: SEER 9 areas (San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, and Atlanta). CRC, colorectal cancer; SEER, Surveillance, Epidemiology, and End Results
40
Secular Trend of CRC in the U.S. in Relation to Folate Status
30 20 10
77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05 20 07 20 09
19
19
75
0
Year of diagnosis
Figure 1. Age-standardized incidence rates of CRC for all ages, all races, and both genders (1975–2009) Note: Cancer sites include invasive cases only unless otherwise noted. Rates are per 100,000 and are age-adjusted to the 2000 U.S. Std population (19 age groups, Census P25–1130). Regression lines are calculated using the Joinpoint Regression Program Version 3.5, April 2011, National Cancer Institute. Data source: SEER 9 areas (San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, and Atlanta). CRC, colorectal cancer; SEER, Surveillance, Epidemiology, and End Results
March 2014
A striking observation is that whites aged Z65 years experienced a sharp reduction in incidence rates in the mid-1980s whereas blacks aged Z65 years did in the late 2000s (Figure 2). These secular trends evoke a hypothesis that adequate folate status during the ages of 45–64 years may confer benefit on CRC risk later in life. As described previously, as the major beneficiaries of the 1973 reform, a substantial proportion of whites probably moved from low to adequate folate status since 1973. If adequate folate status during the ages of 45–64 years truly confers protection against CRC in later life, the white supplement users aged 45–64 years in 1973 should develop less CRC after allowing sufficient time for a long induction period between adequate folate intake and CRC. Thus, a
S68
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72 450 400
Rate per 100,000
350 300 250 200 150 100 50
19
7 19 5 7 19 7 7 19 9 8 19 1 8 19 3 8 19 5 8 19 7 8 19 9 9 19 1 9 19 3 9 19 5 9 19 7 9 20 9 0 20 1 0 20 3 0 20 5 0 20 7 09
0
Year of diagnosis/death Incidence rate
Death rate
Figure 3. Incidence/death rates of CRC for whites aged Z65 (1975–2009)
Note: Cancer sites include invasive cases only unless otherwise noted. Rates are per 100,000. Regression lines are calculated using the Joinpoint Regression Program Version 3.5, April 2011, National Cancer Institute. Data source for incidence rates: SEER 9 areas (San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, and Atlanta). Data source for death rates: U.S. Mortality Files, National Center for Health Statistics, CDC. CRC, colorectal cancer; SEER, Surveillance, Epidemiology, and End Results
reduction in incidence rates of CRC would be predicted among whites aged Z65 years in the mid-1980s, which is precisely what was observed in national data (Figure 2A). Likewise, blacks had the greatest potential to benefit from the fortification in the mid-1990s because of their lowest pre-fortification folate status. Thus, if the hypothesis is true, then when blacks aged 45–64 years during the fortification era became older in years 2005–2010, beginning roughly a decade after the practical fortification onset (1996), they should have a reduced risk of CRC. Indeed, a sharp reduction in incidence rates in blacks aged Z65 years was observed over this time period (Figure 2B). Although the data described are compatible with a reduction in CRC for both whites and blacks after a decade of respective improved folate status, another important observation is that only whites, particularly those aged Z65 years, experienced an abrupt reversal of downward trends in incidence rates with the initiation of the fortification (Figure 2A). Some raised the concern that mandatory folic acid fortification might be the cause of the increase in incidence rates.7,8 However, this explanation7,8 is inconsistent with several lines of epidemiologic evidence that suggests a reduced CRC risk associated with folate intake after a long induction period but does not support a corresponding immediate increase in risk, even among supplement users.10,11,19,20 First, a pooled analysis of 13 prospective cohort studies with 7–20 years of follow-up showed that high folate
intake was significantly associated with a lower risk of colon cancer (pooled rate ratio [RR] for Z560 vs o240 μg/day=0.87, 95% CI=0.78, 0.98; ptrend=0.009).19 Further, in a large cohort study conducted after initiation of the mandatory fortification (1999–2007), higher consumption of total folate intake was not associated with an increased risk of CRC during the first few years (1999– 2001), and a significant inverse association was observed during the 2001–2007 interval.20 Second, even in the only other human evidence (other than the secular trend data) that supported an enhancement of CRC with folate, daily supplementation with 1000 μg folic acid (in addition to regular diet in the fortification era) elevated the risk of advanced lesions only after 3–5 years of follow-up but not during the first 3 years.21 Of note, after the 1996 fortification, the percentage of individuals with total folate intake over 1000 μg/day remained less than 5%.15 Third, according to a dose–response meta-analysis including ten nested case–control studies, there was no evidence of an increased risk of CRC up to circulating folate level of 73 nmol/L (summary RR for each 10 nmol/L of circulating folate¼0.96, 95% CI¼0.91, 1.02; pheterogeneity¼0.58).22 Following folic acid fortification in the U.S., the mean circulating folate level increased from 19.5 nmol/L to 37.2 nmol/L in non-users of B vitamins and from 30.8 to 40.6 in users of B vitamin supplements.23 Finally, if the temporary increase in CRC incidence rates had been truly caused by excess folate intake resulting from the fortification, one would expect a similar increase among whites soon after FDA’s regulatory change in 1973. However, Figure 2A suggests no such indication. Although these data10,11,19,20 are reassuring against an increase of CRC following folic acid fortification or supplementation in the general population, it has been argued that there could be subgroups at high risk, particularly those with advanced adenomas or earlystage cancers.24,25 However, in any folic acid intervention trials, most of the cancers diagnosed in the first years were likely to originate from covert advanced adenomas or latent cancers. Because of randomization, the proportion of individuals with such neoplasms should have been equal in the intervention and placebo groups at baseline. Thus, if folic acid supplementation promotes the growth of pre-existing neoplasms, an increased risk of CRC should have been observed in trials. However, in a recent meta-analysis of 13 randomized, placebo-controlled trials, folic acid supplementation at doses (500–40,000 μg/day) considerably higher than those from fortification did not substantially increase the risk of CRC during the first 5 years of intervention (pooled RR=1.07, 95% CI=0.83, 1.37; p=0.49).26 Albeit underpowered to www.ajpmonline.org
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
S69
B. White women
A. White men 500
350
450 300
350
Rate per 100,000
Rate per 100,000
400
300 250 200 150
250 200 150 100
100 50
50
0
19 7 19 5 77 19 7 19 9 81 19 8 19 3 8 19 5 87 19 8 19 9 91 19 9 19 3 9 19 5 97 19 9 20 9 0 20 1 03 20 0 20 5 07 20 09
19 7 19 5 77 19 7 19 9 81 19 8 19 3 85 19 8 19 7 8 19 9 91 19 9 19 3 9 19 5 97 19 9 20 9 01 20 0 20 3 05 20 0 20 7 09
0
Year of diagnosis/death
450
Year of diagnosis/death
D. Black women
C. Black men 350
400
300 Rate per 100,000
Rate per 100,000
350 300 250 200 150
250 200 150 100
100
50
50
0
19 7 19 5 77 19 7 19 9 81 19 8 19 3 8 19 5 8 19 7 89 19 9 19 1 9 19 3 95 19 9 19 7 99 20 0 20 1 03 20 0 20 5 0 20 7 09
19 7 19 5 7 19 7 79 19 8 19 1 8 19 3 85 19 8 19 7 8 19 9 9 19 1 93 19 9 19 5 97 19 9 20 9 0 20 1 0 20 3 05 20 0 20 7 09
0
Year of diagnosis/death
Year of diagnosis/death
Incidence rate for ages (years):
20–49
Death rate for ages (years):
≥65
50–64
≥65
Figure 4. Age-specific incidence/death rates of CRC by race and gender (1975–2009): (A) white men; (B) white women; (C) black men; (D) black women Note: Cancer sites include invasive cases only unless otherwise noted. Rates are per 100,000. Regression lines are calculated using the Joinpoint Regression Program Version 3.5, April 2011, National Cancer Institute. Data source for incidence rates: SEER 9 areas (San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, and Atlanta). Data source for death rates: U.S. Mortality Files, National Center for Health Statistics, CDC. CRC, colorectal cancer; SEER, Surveillance, Epidemiology, and End Results
detect moderate associations, even in its subgroup analysis including three trials whose participants had a prior history of adenomas and thus a higher risk for CRC, no evidence for an increased risk of CRC associated with folic acid supplementation was apparent (RR¼0.76, 95% CI¼0.32, 1.80; p¼0.53).26 Although a potential role of excess folate in progressing pre-existing subclinical neoplasms could not be completely dismissed, it is further weakened by the plausibility of an alternative hypothesis that folate might have a diminished or minimal protective effect on already established legions.20,27 For instance, in a large cohort study, a significant protective association between total folate intake and CRC risk became more pronounced March 2014
when the analysis was limited to those with history of endoscopy and, thus, less likely to harbor advanced adenomas or early-stage cancers.20
Secular Trend of CRC in the U.S. in Relation to CRC Screening Practices Although the arguments summarized indicate that folic acid fortification did not account for the bump in CRC incidence rates among whites aged Z65 years in the later 1990s, it is important to consider alternative explanations. Several observations are consistent with attributing the sudden increase in incidence rates to a screening effect. First, the observed pattern of a temporary increase in incidence rates followed by a return back to the
S70
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
downward trends while death rates kept decreasing (Figure 3) resembles the distinctive characteristic of the screening bump, explained previously. Furthermore, the specificity of the bump among whites, with a more pronounced effect among white men aged Z65 years, strengthens the evidence for the screening effect (Figure 4). Considering that meaningful facilitators to CRC screening include white race and male gender28 and that age is an independent risk factor for CRC,5 if the momentary increase in incidence rates had been truly due to the screening effect, the bump would have been most evident among elderly white men, which is consistent with national data (Figure 4A). Of note, effective since January 1, 1998, was the expansion of Medicare coverage to include fecal occult blood testing (FOBT), sigmoidoscopy, and colonoscopy for “screening” purposes among high-risk individuals.29 Given that health insurance coverage is an important facilitator for CRC screening28 and that the beneficiaries of Medicare are disproportionally whites and elderly over 65 years, the change in Medicare policy should have most strongly affected older whites, which is observed in Figure 4 where only whites aged Z65 years showed the screening bump in the late 1990s. In light of the previous explanation that the screening bump occurs immediately on the initial increase in screening rates, one might argue that mismatch in time between the beginning of the observed bump in 1995 and the change in Medicare policy in 1998 counters the screening effect. However, some evidence supports that the start of increasing screening rates dates back to 1995. Fenton et al.29 examined the annual trends of FOBT, sigmoidoscopy, and colonoscopy from 1995 to 2003 among annual 5% random samples of the Medicare beneficiaries residing in the SEER areas, from which the data for the secular trends of CRC are generated. The study showed that since 1995, colonoscopy started to replace sigmoidoscopy, especially among whites, and that each year the cumulative increase in colonoscopy was greater than the cumulative decrease in sigmoidoscopy, resulting in a net annual increase in rates of endoscopic tests for CRC starting from 1995. This is also consistent with the data from the Health Professionals Follow-up Study, in which majority of the study population is white and the rates of colonoscopy for screening purposes started to increase from 1995 (ELG, unpublished observations, 2013). Thus, both the rising rates in screening colonoscopy and the temporary increase in CRC incidence rates beginning in 1995 supports the screening effect. Although available data may not be precise enough to designate definitively when the rise in colonoscopy usage in the mid-1990s would have been reflected in national
incidence rates of CRC, a screening bump should have been evident at some point over this time period given the substantial increase in endoscopic screening. Because the observed bump in the late 1990s is the only notable increase in incidence, it is likely due to the screening effect. Subsequently, as explained previously, a reduction in CRC is expected to follow the bump as the removal of precancerous adenomas detected by screening leads to the primary prevention of CRC, which is precisely what was observed in national data (Figure 2A). Thus, for whites aged Z65 years, although improved folate status by FDA’s regulatory change in 1973 may have contributed to the decrease in CRC incidence rates between mid1980s and mid-1990s, it is the expansion of screening endoscopy that is likely to explain a large proportion of the decrease since the 2000s.
Conclusion and Future Directions Examination of the secular trend of a disease within one country provides an opportunity to investigate the effects of nongenetic factors on the disease. Because of inherent methodologic limitations such as use of group-level data rather than individual-level information, lack of confounding control, and failure to account for the influence of changing prevalence of other risk factors over time, this type of analysis precludes causal inference. Nonetheless, by integrating the current state of knowledge on subject matters, examination of the temporal association between a nongenetic factor and a disease can engender a meaningful hypothesis that could suggest a direction for future research. In this investigation on the U.S. secular trend of CRC in relation to changing folate status, we simultaneously accounted for a long time lag between change in folate status and change in CRC incidence rates, two major changes in FDA’s regulations that affected nationwide folate intake, differential impacts of these policy changes by race, and screening trends. This approach helped to generate a hypothesis that adequate folate status during the ages of 45–64 years may confer benefit on CRC risk in later life, and to alleviate the enduring concern that the momentary, abrupt increase in CRC incidence rates in the mid-1990s might have resulted from excessive folate intake caused by folic acid fortification. Surely, the proposed hypothesis is not conclusive because of the aforementioned methodologic limitations. However, in the U.S., despite the fact that the prevalence of major CRC risk factors such as the Western dietary pattern and obesity has been undoubtedly increasing, the overall incidence rates of CRC have continued to decline. Although protective agents such as calcium, aspirin, hormone replacement therapy, and screening endoscopy www.ajpmonline.org
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
and nationwide reduction in smoking could have contributed to the downward trend of CRC incidence rates, the high prevalence of supplement users and mandatory folic acid fortification provides a unique environment to the U.S. Thus, folate is more likely to explain the distinctiveness of the U.S. being the only country in the world where CRC incidence rates are decreasing. At present, long-term cohort studies and their pooled analysis provide compelling evidence for the benefit of adequate folate intake on CRC.10,11,19 However, results from RCTs are conflicting, and meta-analyses of trials found no significant effect of folic acid supplementation on CRC.22,26 In general, when assessing the current state of evidence for the causal relationship between an exposure and an outcome, more weight is given to results from RCTs because of their methodologic strengths. However, assuming a long induction period for folate to exert its full effect on CRC risk, most of the existing trials22,26 are limited by their short duration of follow-up as to provide a definite conclusion. It is entirely possible that an important benefit of folic acid supplementation on CRC could manifest on an extended follow-up that is sufficient enough to allow for the long induction period. Furthermore, because trial participants are mostly health conscious and well nourished and may already have had adequate folate status, trials targeting individuals with suboptimal folate status might show a significant benefit of folate intake on CRC risk. Indeed, in a recent RCT among Chinese subjects with a relatively low baseline folate status, 1000 μg/day supplementation of folic acid significantly reduced the risk of incident adenoma (RR=0.49, 95% CI=0.37, 0.63; po0.01).30 Finally, most participants in the existing trials22,26 were elderly (mean baseline age 460 years) and, thus, they were not able to address the question of whether adequate folate intake during the ages of 45–64 years confers a benefit on CRC later in life. In conclusion, a number of observations suggest that the increase in CRC incidence rates in the later 1990s in the U.S. is unlikely to be due to folic acid fortification. In contrast, when assuming a time lag of a decade or longer to see a benefit on CRC, as suggested by some studies,10,11 folate appears to be one of the most promising factors that could contribute to the unique downward trend of CRC incidence rates in the U.S. More studies are needed to provide definitive evidence to address the potential benefits and risks of folate on CRC incidence rates and mortality. In particular, identification of the dose– response relationship between total folate intake and CRC risk across a wide range of folate intake over a sufficient follow-up during the fortification era would help answer whether there is a range of folate intake that could optimally protect against CRC. A pooled study of March 2014
S71
individuals with history of adenomas recruited from long-established cohorts is warranted to address controversy surrounding a potential role of excess folate in promoting pre-existing neoplasms with adequate statistical power. Meanwhile, given that the prevalence of folate deficiency is o1% during post-fortification in the U.S.,31 it would be most reasonable that individuals aged 45–64 years, except for those at risk for inadequate folate status (alcohol abusers, nutrient malabsorbers) target other risk factors2 and follow recommended screening schedule for the prevention of CRC. Other countries, especially those with a low folate status, may consider implementing nationwide folic acid fortification to harness diverse health benefits of achieving adequate folate status. This research was supported by Harvard School of Public Health Department of Nutrition. The publication of this supplement was made possible through the CDC and the Association for Prevention Teaching and Research (APTR) Cooperative Agreement No. 1 U360E000005-01. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC or the APTR. No financial disclosures were reported by the authors of this paper.
References 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer;127(12):2893–2917. 2. Slattery ML, Edwards SL, Boucher KM, Anderson K, Caan BJ. Lifestyle and colon cancer: an assessment of factors associated with risk. Am J Epidemiol 1999;150(8):869–77. 3. Center MM, Jemal A, Ward E. International trends in colorectal cancer incidence rates. Cancer Epidemiol Biomarkers Prev 2009;18(6): 1688–94. 4. Edwards BK, Ward E, Kohler BA, et al. Annual report to the nation on the status of cancer, 1975–2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010;116(3):544–73. 5. Giovannucci E, Wu K. Cancers of the colon and rectum. In: Cancer epidemiology and prevention, 3rd ed. New York: Oxford University Press, 2006:809–29. 6. U.S. Food and Drug Administration. Statement of general policy or interpretation. Subchapter B—Food and food products, Part 121— Food additives. Fed Register 1973;38:20725–6. 7. Mason JB, Dickstein A, Jacques PF, et al. A temporal association between folic acid fortification and an increase in colorectal cancer rates may be illuminating important biological principles: a hypothesis. Cancer Epidemiol Biomarkers Prev 2007;16(7):1325–9. 8. Osterhues A, Holzgreve W, Michels KB. Shall we put the world on folate? Lancet 2009;374(9694):959–61. 9. McGlynn AP, Wasson GR, O’Reilly SL, et al. Low colonocyte folate is associated with uracil misincorporation and global DNA hypomethylation in human colorectum. J Nutr 2013;143(1):27–33.
S72
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
10. Giovannucci E, Stampfer MJ, Colditz GA, et al. Multivitamin use, folate, and colon cancer in women in the Nurses0 Health Study. Ann Intern Med 1998;129(7):517–24. 11. Lee JE, Willett WC, Fuchs CS, et al. Folate intake and risk of colorectal cancer and adenoma: modification by time. Am J Clin Nutr 2011;93(4): 817–25. 12. Winawer SJ. Natural history of colorectal cancer. Am J Med 1999;106 (1A):3S–6S; discussion 50S–51S. 13. CDC. Folic acid recommendations. 2012 January 13. www.cdc.gov/ ncbddd/folicacid/recommendations.html. 14. Briefel RR, Johnson CL. Secular trends in dietary intake in the U.S. Annu Rev Nutr 2004;24:401–31. 15. Bentley TG, Willett WC, Weinstein MC, Kuntz KM. Population-level changes in folate intake by age, gender, and race/ethnicity after folic acid fortification. Am J Public Health 2006;96(11):2040–7. 16. Wu K, Platz EA, Willett WC, et al. A randomized trial on folic acid supplementation and risk of recurrent colorectal adenoma. Am J Clin Nutr 2009;90(6):1623–31. 17. Gibson TM, Weinstein SJ, Pfeiffer RM, et al. Pre- and postfortification intake of folate and risk of colorectal cancer in a large prospective cohort study in the U.S. Am J Clin Nutr 2011;94(4):1053–62. 18. Schoen RE, Pinsky PF, Weissfeld JL, et al. Colorectal-cancer incidence and mortality with screening flexible sigmoidoscopy. N Engl J Med 2012;366(25):2345–57. 19. Kim DH, Smith-Warner SA, Spiegelman D, et al. Pooled analyses of 13 prospective cohort studies on folate intake and colon cancer. Cancer Causes Control 2010;21(11):1919–30. 20. Stevens VL, McCullough ML, Sun J, Jacobs EJ, Campbell PT, Gapstur SM. High levels of folate from supplements and fortification are not associated with increased risk of colorectal cancer. Gastroenterology 2011;141(1):98–105, 105 e1.
21. Cole BF, Baron JA, Sandler RS, et al. Folic acid for the prevention of colorectal adenomas: a randomized clinical trial. JAMA 2007;297(21):2351–9. 22. Qin X, Cui Y, Shen L, et al. Folic acid supplementation and cancer risk: a meta-analysis of randomized controlled trials. Int J Cancer 2013;133 (5):1033–41. 23. Kalmbach RD, Choumenkovitch SF, Troen AM, D’Agostino R, Jacques PF, Selhub J. Circulating folic acid in plasma: relation to folic acid fortification. Am J Clin Nutr 2008;88(3):763–8. 24. Mason JB. Folate consumption and cancer risk: a confirmation and some reassurance, but we’re not out of the woods quite yet. Am J Clin Nutr 2011;94(4):965–6. 25. Ulrich CM, Potter JD. Folate and cancer—timing is everything. JAMA 2007;297(21):2408–9. 26. Vollset SE, Clarke R, Lewington S, et al. Effects of folic acid supplementation on overall and site-specific cancer incidence during the randomised trials: meta-analyses of data on 50,000 individuals. Lancet 2013;381(9871):1029–36. 27. Lin YW, Wang JL, Chen HM, et al. Folic acid supplementary reduce the incidence of adenocarcinoma in a mouse model of colorectal cancer: microarray gene expression profile. J Exp Clin Cancer Res 2011;30:116. 28. Guessous I, Dash C, Lapin P, Doroshenk M, Smith RA, Klabunde CN. Colorectal cancer screening barriers and facilitators in older persons. Prev Med 2010;50(1–2):3–10. 29. Fenton JJ, Cai Y, Green P, Beckett LA, Franks P, Baldwin LM. Trends in colorectal cancer testing among Medicare subpopulations. Am J Prev Med 2008;35(3):194–202. 30. Gao QY, Chen HM, Chen YX, et al. Folic Acid prevents the initial occurrence of sporadic colorectal adenoma in Chinese older than 50 years of age: a randomized clinical trial. Cancer Prev Res (Phila) 2013;6(7):744–52. 31. CDC. Second national report on biochemical indicators of diet and nutrition in the U.S. population. CDC, 2012. www.cdc.gov/nutritionreport/.
www.ajpmonline.org
Introduction
C
olorectal cancer (CRC), as the third most commonly diagnosed cancer and the fourth-leading cause of cancer death worldwide,1 affects predominantly developed countries with the Western lifestyle.2 Although incidence rates of CRC have generally increased in developing countries and have stabilized in developed countries in recent decades, notably, the U.S. is the only country experiencing a decline in CRC incidence rates.3 This decline, if not all due to screening, indicates the presence of increasing prevalence of protective factors specific to the U.S. One study that compared CRC incidence rates observed with those predicted by MISCAN-COLON model, a simulation model that estimated secular trends of CRC if changes in the prevalence of risk factors or screening practice had not From the Department of Nutrition and Epidemiology (Keum, Giovannucci), Harvard School of Public Health; and the Channing Division of Network Medicine (Giovannucci), Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts Address correspondence to: Edward L. Giovannucci, MD, ScD, Department of Nutrition and Epidemiology, Harvard School of Public Health, 665 Huntington Avenue, Boston MA 02115. E-mail: [email protected]. edu. 0749-3797/$36.00 http://dx.doi.org/10.1016/j.amepre.2013.10.025
& 2014 American Journal of Preventive Medicine
occurred, showed that about half the decline was accounted for by changes in risk factors and the rest by screening.4 One potential factor is folate (vitamin B9 found naturally in foods), which is suggested to protect against CRC by preventing aberrations in DNA synthesis (DNA uracil misincorporation) and irregularities in DNA methylation (DNA hypomethylation) that promote colorectal carcinogenesis.5 Folate status has increased substantially in the U.S. resulting from a change in U.S. Food and Drug Administration (FDA)’s regulation in 1973 on the dose of folic acid (the synthetic form of folate) allowed in supplements6 and from the fortification (140 μg folic acid/100 g) of flour and grains with folic acid in the 1990s.7 Of note, although the fortification became mandatory in 1998, it virtually started from 1996 because of voluntary fortification by food industries immediately after the release of the new regulation in 1996.7 Past analyses that overlaid the secular trend of CRC with changes in national folate intake over time created a concern that excessive folate intake might promote CRC rather than preventing it.7,8 However, these analyses7,8 are limited by their failure to simultaneously address the following four critical factors: (1) a long induction period between improved folate status and potential protection of CRC; (2) a change in FDA’s regulation in 1973 on the
Published by Elsevier Inc.
Am J Prev Med 2014;46(3S1):S65–S72 S65
S66
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
dose of folic acid allowed in supplements; (3) differential impacts of 1973 regulatory change and 1990s mandatory fortification by race; and (4) changes in CRC screening over time in the U.S. Thus, this article will examine the CRC trend in the U.S. in relation to folate intake by simultaneously accounting for the aforementioned four factors and consider the potential benefits as well as risks of increased folate status on CRC.
Long Induction Period A long induction period deserves special attention in the study of CRC because the progression from normal cells of the colon and rectum, to small adenomas, to large adenomas, and finally to adenocarcinomas, occurs over a long duration, up to 30–40 years.5 The presence of a long induction period between adequate folate status and reduced risk of CRC is evident from both molecular mechanistic9 and epidemiologic studies.10,11 In a study that measured total folate concentration in colonocytes of adenoma patients (cases) and individuals without adenomas (controls), there was a significant decrease in colonocyte folate levels from the normal tissue of controls, to the normal tissue of cases adjacent to adenomas, and to the adenoma itself.9 The lower folate levels corresponded to increases in uracil misincorporation and DNA hypomethylation even in normal-appearing, non-neoplastic tissue of cases.9 As it takes at least 10 years in general to progress from non-neoplastic tissue to cancer diagnosis,12 the presence of such genetic alterations in non-neoplastic cells in the proximal field of adenomas implies that macroscopically undetectable abnormalities involving folate deficiency occur at least a decade before CRC is apparent. Epidemiologic evidence also supports a substantial time lag of at least 12–16 years between folate intake and CRC risk.10,11 Such data suggest that a protective effect of increasing folate intake would be reflected in incidence rates of CRC only after allowing for the estimated minimum induction period of 12–16 years.
Change in FDA Regulation on Dose of Folic Acid Although focus has been typically on mandatory folic acid fortification in the 1990s, a marked increase in folic acid intake dates back to 1973 when the FDA increased the daily maximum dose of folic acid allowed in supplements and foods from 100 to 400 μg6 (a recommended dose for women of childbearing age to prevent birth defects).13 Although its impact was not widespread, approximately one quarter of the U.S population using vitamin supplements in 1973 experienced a significant increase in folic acid intake10 along with an additional rise from fortified breakfast cereals. Yet, no previous
study to our knowledge has considered how the increase in folic acid consumption in 1973 relates to population incidence and death rates of CRC. More complete pictures would be obtained if the secular trend of CRC is examined in conjunction with both FDA’s regulatory change in 1973 and mandatory fortification in the 1990s while accounting for a long induction period.
Differential Impacts of 1973 Regulatory Change and 1990's Mandatory Fortification by Race In consequence of the 1973 change, the major group of individuals whose folate status was expected to improve was supplement users. According to the National Health and Nutrition Examination Survey I (1971–1975), the prevalence of supplement use among the adults aged 20– 74 years was 27.5% in men and 37.9% in women, and whites were the major users.14 Thus, any benefit associated with an improved folate status would be more pronounced among whites. Although the fortification policy in the 1990s led to a population-level increase in total folate intake by approximately 100 μg/day, with the largest increase for whites,15 pre-fortification daily total folate intake of adults aged 45 to 64 years was lowest in blacks (242 μg/day in men and 206 μg/day in women).15 Given considerable evidence that additional folate intake results in the greatest reduction in CRC risk among individuals at low folate status at baseline,10,16,17 blacks had the greatest potential to benefit from the fortification. In summary, variations by race in the improvement of folate status predict that the impact of respective folic acid policy on CRC risk would differ by race as well. This emphasizes the critical need for investigating the temporal association between folate intake and CRC by race.
Changes in CRC Screening Over Time in the U.S. Understanding how screening influences incidence rates of CRC over time is crucial to teasing out the effect of screening from that of folate intake. A landmark trial18 that evaluated the effect of sigmoidoscopy showed that approximately for the first 3 years since the baseline randomization, incidence rates were higher in the screening group than in the usual care group because of an increased detection of asymptomatic CRC by screening. However, after Year 3, incidence rates became and remained lower in the screening group presumably because the removal of precancerous adenomas detected by screening led to the primary prevention of CRC. This competing reduction of CRC incidence offset and then outweighed the initial rise due to an increased detection. Taken together, it can be inferred that the initial increase in screening rates is immediately followed www.ajpmonline.org
S67
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
70
Rate per 100,000
60 50
Rate per 100,000
350 300 250 200 150 100 50 19 7 19 5 77 19 7 19 9 8 19 1 83 19 8 19 5 87 19 8 19 9 91 19 9 19 3 95 19 9 19 7 99 20 0 20 1 03 20 0 20 5 07 20 09
0
Year of diagnosis
B. Blacks
450 400 350 300 250 200 150 100 50 0
7 19 5 7 19 7 79 19 8 19 1 8 19 3 85 19 8 19 7 8 19 9 9 19 1 93 19 9 19 5 9 19 7 99 20 0 20 1 0 20 3 05 20 0 20 7 09
Based on data from the SEER, the overall secular trend of CRC in the U.S. was plotted. The trend was further plotted by age, race, and gender. Figure 1 presents the overall secular trend, which shows a sudden increase in CRC incidence rates in the mid-1990s. This is the figure that raised the concern that excess folate intake due to the nationwide fortification might promote colorectal carcinogenesis.7,8 However, Figures 2–4 show substantial heterogeneity in trends by subgroups of age, race, and gender, which underscores that interpretation of the temporal relationship deserves more delicate explanations.
400
19
Procedures
A. Whites
450
Rate per 100,000
by a momentary increase in incidence rates without a time lag for the first few years. Of note, the initial rise in incidence rates in the screening group did not translate into higher death rates. Mortality, as a function of both incidence and prognosis, can be reduced by decreasing incidence or improving prognosis. Screening reduces mortality through both mechanisms: the detection and removal of precancerous lesions lead to the primary prevention of CRC and, thus, lowers incidence; an early detection of asymptomatic CRC improves prognosis. Hence, the decreasing trends in mortality in spite of a momentary rise in incidence rates, which we termed as the screening bump, serve as the characteristic feature that distinguishes the effect of screening from the effects of risk factors that elevate both incidence rates and death rates in most cases.
Year of diagnosis Age (years):
20–49
50–64
≥65
Figure 2. Age-specific incidence rates of CRC for both genders by race (1975–2009): (A) whites, (B) blacks Note: Cancer sites include invasive cases only unless otherwise noted. Rates are per 100,000. Regression lines are calculated using the Joinpoint Regression Program Version 3.5, April 2011, National Cancer Institute. Data source: SEER 9 areas (San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, and Atlanta). CRC, colorectal cancer; SEER, Surveillance, Epidemiology, and End Results
40
Secular Trend of CRC in the U.S. in Relation to Folate Status
30 20 10
77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05 20 07 20 09
19
19
75
0
Year of diagnosis
Figure 1. Age-standardized incidence rates of CRC for all ages, all races, and both genders (1975–2009) Note: Cancer sites include invasive cases only unless otherwise noted. Rates are per 100,000 and are age-adjusted to the 2000 U.S. Std population (19 age groups, Census P25–1130). Regression lines are calculated using the Joinpoint Regression Program Version 3.5, April 2011, National Cancer Institute. Data source: SEER 9 areas (San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, and Atlanta). CRC, colorectal cancer; SEER, Surveillance, Epidemiology, and End Results
March 2014
A striking observation is that whites aged Z65 years experienced a sharp reduction in incidence rates in the mid-1980s whereas blacks aged Z65 years did in the late 2000s (Figure 2). These secular trends evoke a hypothesis that adequate folate status during the ages of 45–64 years may confer benefit on CRC risk later in life. As described previously, as the major beneficiaries of the 1973 reform, a substantial proportion of whites probably moved from low to adequate folate status since 1973. If adequate folate status during the ages of 45–64 years truly confers protection against CRC in later life, the white supplement users aged 45–64 years in 1973 should develop less CRC after allowing sufficient time for a long induction period between adequate folate intake and CRC. Thus, a
S68
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72 450 400
Rate per 100,000
350 300 250 200 150 100 50
19
7 19 5 7 19 7 7 19 9 8 19 1 8 19 3 8 19 5 8 19 7 8 19 9 9 19 1 9 19 3 9 19 5 9 19 7 9 20 9 0 20 1 0 20 3 0 20 5 0 20 7 09
0
Year of diagnosis/death Incidence rate
Death rate
Figure 3. Incidence/death rates of CRC for whites aged Z65 (1975–2009)
Note: Cancer sites include invasive cases only unless otherwise noted. Rates are per 100,000. Regression lines are calculated using the Joinpoint Regression Program Version 3.5, April 2011, National Cancer Institute. Data source for incidence rates: SEER 9 areas (San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, and Atlanta). Data source for death rates: U.S. Mortality Files, National Center for Health Statistics, CDC. CRC, colorectal cancer; SEER, Surveillance, Epidemiology, and End Results
reduction in incidence rates of CRC would be predicted among whites aged Z65 years in the mid-1980s, which is precisely what was observed in national data (Figure 2A). Likewise, blacks had the greatest potential to benefit from the fortification in the mid-1990s because of their lowest pre-fortification folate status. Thus, if the hypothesis is true, then when blacks aged 45–64 years during the fortification era became older in years 2005–2010, beginning roughly a decade after the practical fortification onset (1996), they should have a reduced risk of CRC. Indeed, a sharp reduction in incidence rates in blacks aged Z65 years was observed over this time period (Figure 2B). Although the data described are compatible with a reduction in CRC for both whites and blacks after a decade of respective improved folate status, another important observation is that only whites, particularly those aged Z65 years, experienced an abrupt reversal of downward trends in incidence rates with the initiation of the fortification (Figure 2A). Some raised the concern that mandatory folic acid fortification might be the cause of the increase in incidence rates.7,8 However, this explanation7,8 is inconsistent with several lines of epidemiologic evidence that suggests a reduced CRC risk associated with folate intake after a long induction period but does not support a corresponding immediate increase in risk, even among supplement users.10,11,19,20 First, a pooled analysis of 13 prospective cohort studies with 7–20 years of follow-up showed that high folate
intake was significantly associated with a lower risk of colon cancer (pooled rate ratio [RR] for Z560 vs o240 μg/day=0.87, 95% CI=0.78, 0.98; ptrend=0.009).19 Further, in a large cohort study conducted after initiation of the mandatory fortification (1999–2007), higher consumption of total folate intake was not associated with an increased risk of CRC during the first few years (1999– 2001), and a significant inverse association was observed during the 2001–2007 interval.20 Second, even in the only other human evidence (other than the secular trend data) that supported an enhancement of CRC with folate, daily supplementation with 1000 μg folic acid (in addition to regular diet in the fortification era) elevated the risk of advanced lesions only after 3–5 years of follow-up but not during the first 3 years.21 Of note, after the 1996 fortification, the percentage of individuals with total folate intake over 1000 μg/day remained less than 5%.15 Third, according to a dose–response meta-analysis including ten nested case–control studies, there was no evidence of an increased risk of CRC up to circulating folate level of 73 nmol/L (summary RR for each 10 nmol/L of circulating folate¼0.96, 95% CI¼0.91, 1.02; pheterogeneity¼0.58).22 Following folic acid fortification in the U.S., the mean circulating folate level increased from 19.5 nmol/L to 37.2 nmol/L in non-users of B vitamins and from 30.8 to 40.6 in users of B vitamin supplements.23 Finally, if the temporary increase in CRC incidence rates had been truly caused by excess folate intake resulting from the fortification, one would expect a similar increase among whites soon after FDA’s regulatory change in 1973. However, Figure 2A suggests no such indication. Although these data10,11,19,20 are reassuring against an increase of CRC following folic acid fortification or supplementation in the general population, it has been argued that there could be subgroups at high risk, particularly those with advanced adenomas or earlystage cancers.24,25 However, in any folic acid intervention trials, most of the cancers diagnosed in the first years were likely to originate from covert advanced adenomas or latent cancers. Because of randomization, the proportion of individuals with such neoplasms should have been equal in the intervention and placebo groups at baseline. Thus, if folic acid supplementation promotes the growth of pre-existing neoplasms, an increased risk of CRC should have been observed in trials. However, in a recent meta-analysis of 13 randomized, placebo-controlled trials, folic acid supplementation at doses (500–40,000 μg/day) considerably higher than those from fortification did not substantially increase the risk of CRC during the first 5 years of intervention (pooled RR=1.07, 95% CI=0.83, 1.37; p=0.49).26 Albeit underpowered to www.ajpmonline.org
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
S69
B. White women
A. White men 500
350
450 300
350
Rate per 100,000
Rate per 100,000
400
300 250 200 150
250 200 150 100
100 50
50
0
19 7 19 5 77 19 7 19 9 81 19 8 19 3 8 19 5 87 19 8 19 9 91 19 9 19 3 9 19 5 97 19 9 20 9 0 20 1 03 20 0 20 5 07 20 09
19 7 19 5 77 19 7 19 9 81 19 8 19 3 85 19 8 19 7 8 19 9 91 19 9 19 3 9 19 5 97 19 9 20 9 01 20 0 20 3 05 20 0 20 7 09
0
Year of diagnosis/death
450
Year of diagnosis/death
D. Black women
C. Black men 350
400
300 Rate per 100,000
Rate per 100,000
350 300 250 200 150
250 200 150 100
100
50
50
0
19 7 19 5 77 19 7 19 9 81 19 8 19 3 8 19 5 8 19 7 89 19 9 19 1 9 19 3 95 19 9 19 7 99 20 0 20 1 03 20 0 20 5 0 20 7 09
19 7 19 5 7 19 7 79 19 8 19 1 8 19 3 85 19 8 19 7 8 19 9 9 19 1 93 19 9 19 5 97 19 9 20 9 0 20 1 0 20 3 05 20 0 20 7 09
0
Year of diagnosis/death
Year of diagnosis/death
Incidence rate for ages (years):
20–49
Death rate for ages (years):
≥65
50–64
≥65
Figure 4. Age-specific incidence/death rates of CRC by race and gender (1975–2009): (A) white men; (B) white women; (C) black men; (D) black women Note: Cancer sites include invasive cases only unless otherwise noted. Rates are per 100,000. Regression lines are calculated using the Joinpoint Regression Program Version 3.5, April 2011, National Cancer Institute. Data source for incidence rates: SEER 9 areas (San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, and Atlanta). Data source for death rates: U.S. Mortality Files, National Center for Health Statistics, CDC. CRC, colorectal cancer; SEER, Surveillance, Epidemiology, and End Results
detect moderate associations, even in its subgroup analysis including three trials whose participants had a prior history of adenomas and thus a higher risk for CRC, no evidence for an increased risk of CRC associated with folic acid supplementation was apparent (RR¼0.76, 95% CI¼0.32, 1.80; p¼0.53).26 Although a potential role of excess folate in progressing pre-existing subclinical neoplasms could not be completely dismissed, it is further weakened by the plausibility of an alternative hypothesis that folate might have a diminished or minimal protective effect on already established legions.20,27 For instance, in a large cohort study, a significant protective association between total folate intake and CRC risk became more pronounced March 2014
when the analysis was limited to those with history of endoscopy and, thus, less likely to harbor advanced adenomas or early-stage cancers.20
Secular Trend of CRC in the U.S. in Relation to CRC Screening Practices Although the arguments summarized indicate that folic acid fortification did not account for the bump in CRC incidence rates among whites aged Z65 years in the later 1990s, it is important to consider alternative explanations. Several observations are consistent with attributing the sudden increase in incidence rates to a screening effect. First, the observed pattern of a temporary increase in incidence rates followed by a return back to the
S70
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
downward trends while death rates kept decreasing (Figure 3) resembles the distinctive characteristic of the screening bump, explained previously. Furthermore, the specificity of the bump among whites, with a more pronounced effect among white men aged Z65 years, strengthens the evidence for the screening effect (Figure 4). Considering that meaningful facilitators to CRC screening include white race and male gender28 and that age is an independent risk factor for CRC,5 if the momentary increase in incidence rates had been truly due to the screening effect, the bump would have been most evident among elderly white men, which is consistent with national data (Figure 4A). Of note, effective since January 1, 1998, was the expansion of Medicare coverage to include fecal occult blood testing (FOBT), sigmoidoscopy, and colonoscopy for “screening” purposes among high-risk individuals.29 Given that health insurance coverage is an important facilitator for CRC screening28 and that the beneficiaries of Medicare are disproportionally whites and elderly over 65 years, the change in Medicare policy should have most strongly affected older whites, which is observed in Figure 4 where only whites aged Z65 years showed the screening bump in the late 1990s. In light of the previous explanation that the screening bump occurs immediately on the initial increase in screening rates, one might argue that mismatch in time between the beginning of the observed bump in 1995 and the change in Medicare policy in 1998 counters the screening effect. However, some evidence supports that the start of increasing screening rates dates back to 1995. Fenton et al.29 examined the annual trends of FOBT, sigmoidoscopy, and colonoscopy from 1995 to 2003 among annual 5% random samples of the Medicare beneficiaries residing in the SEER areas, from which the data for the secular trends of CRC are generated. The study showed that since 1995, colonoscopy started to replace sigmoidoscopy, especially among whites, and that each year the cumulative increase in colonoscopy was greater than the cumulative decrease in sigmoidoscopy, resulting in a net annual increase in rates of endoscopic tests for CRC starting from 1995. This is also consistent with the data from the Health Professionals Follow-up Study, in which majority of the study population is white and the rates of colonoscopy for screening purposes started to increase from 1995 (ELG, unpublished observations, 2013). Thus, both the rising rates in screening colonoscopy and the temporary increase in CRC incidence rates beginning in 1995 supports the screening effect. Although available data may not be precise enough to designate definitively when the rise in colonoscopy usage in the mid-1990s would have been reflected in national
incidence rates of CRC, a screening bump should have been evident at some point over this time period given the substantial increase in endoscopic screening. Because the observed bump in the late 1990s is the only notable increase in incidence, it is likely due to the screening effect. Subsequently, as explained previously, a reduction in CRC is expected to follow the bump as the removal of precancerous adenomas detected by screening leads to the primary prevention of CRC, which is precisely what was observed in national data (Figure 2A). Thus, for whites aged Z65 years, although improved folate status by FDA’s regulatory change in 1973 may have contributed to the decrease in CRC incidence rates between mid1980s and mid-1990s, it is the expansion of screening endoscopy that is likely to explain a large proportion of the decrease since the 2000s.
Conclusion and Future Directions Examination of the secular trend of a disease within one country provides an opportunity to investigate the effects of nongenetic factors on the disease. Because of inherent methodologic limitations such as use of group-level data rather than individual-level information, lack of confounding control, and failure to account for the influence of changing prevalence of other risk factors over time, this type of analysis precludes causal inference. Nonetheless, by integrating the current state of knowledge on subject matters, examination of the temporal association between a nongenetic factor and a disease can engender a meaningful hypothesis that could suggest a direction for future research. In this investigation on the U.S. secular trend of CRC in relation to changing folate status, we simultaneously accounted for a long time lag between change in folate status and change in CRC incidence rates, two major changes in FDA’s regulations that affected nationwide folate intake, differential impacts of these policy changes by race, and screening trends. This approach helped to generate a hypothesis that adequate folate status during the ages of 45–64 years may confer benefit on CRC risk in later life, and to alleviate the enduring concern that the momentary, abrupt increase in CRC incidence rates in the mid-1990s might have resulted from excessive folate intake caused by folic acid fortification. Surely, the proposed hypothesis is not conclusive because of the aforementioned methodologic limitations. However, in the U.S., despite the fact that the prevalence of major CRC risk factors such as the Western dietary pattern and obesity has been undoubtedly increasing, the overall incidence rates of CRC have continued to decline. Although protective agents such as calcium, aspirin, hormone replacement therapy, and screening endoscopy www.ajpmonline.org
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
and nationwide reduction in smoking could have contributed to the downward trend of CRC incidence rates, the high prevalence of supplement users and mandatory folic acid fortification provides a unique environment to the U.S. Thus, folate is more likely to explain the distinctiveness of the U.S. being the only country in the world where CRC incidence rates are decreasing. At present, long-term cohort studies and their pooled analysis provide compelling evidence for the benefit of adequate folate intake on CRC.10,11,19 However, results from RCTs are conflicting, and meta-analyses of trials found no significant effect of folic acid supplementation on CRC.22,26 In general, when assessing the current state of evidence for the causal relationship between an exposure and an outcome, more weight is given to results from RCTs because of their methodologic strengths. However, assuming a long induction period for folate to exert its full effect on CRC risk, most of the existing trials22,26 are limited by their short duration of follow-up as to provide a definite conclusion. It is entirely possible that an important benefit of folic acid supplementation on CRC could manifest on an extended follow-up that is sufficient enough to allow for the long induction period. Furthermore, because trial participants are mostly health conscious and well nourished and may already have had adequate folate status, trials targeting individuals with suboptimal folate status might show a significant benefit of folate intake on CRC risk. Indeed, in a recent RCT among Chinese subjects with a relatively low baseline folate status, 1000 μg/day supplementation of folic acid significantly reduced the risk of incident adenoma (RR=0.49, 95% CI=0.37, 0.63; po0.01).30 Finally, most participants in the existing trials22,26 were elderly (mean baseline age 460 years) and, thus, they were not able to address the question of whether adequate folate intake during the ages of 45–64 years confers a benefit on CRC later in life. In conclusion, a number of observations suggest that the increase in CRC incidence rates in the later 1990s in the U.S. is unlikely to be due to folic acid fortification. In contrast, when assuming a time lag of a decade or longer to see a benefit on CRC, as suggested by some studies,10,11 folate appears to be one of the most promising factors that could contribute to the unique downward trend of CRC incidence rates in the U.S. More studies are needed to provide definitive evidence to address the potential benefits and risks of folate on CRC incidence rates and mortality. In particular, identification of the dose– response relationship between total folate intake and CRC risk across a wide range of folate intake over a sufficient follow-up during the fortification era would help answer whether there is a range of folate intake that could optimally protect against CRC. A pooled study of March 2014
S71
individuals with history of adenomas recruited from long-established cohorts is warranted to address controversy surrounding a potential role of excess folate in promoting pre-existing neoplasms with adequate statistical power. Meanwhile, given that the prevalence of folate deficiency is o1% during post-fortification in the U.S.,31 it would be most reasonable that individuals aged 45–64 years, except for those at risk for inadequate folate status (alcohol abusers, nutrient malabsorbers) target other risk factors2 and follow recommended screening schedule for the prevention of CRC. Other countries, especially those with a low folate status, may consider implementing nationwide folic acid fortification to harness diverse health benefits of achieving adequate folate status. This research was supported by Harvard School of Public Health Department of Nutrition. The publication of this supplement was made possible through the CDC and the Association for Prevention Teaching and Research (APTR) Cooperative Agreement No. 1 U360E000005-01. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC or the APTR. No financial disclosures were reported by the authors of this paper.
References 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer;127(12):2893–2917. 2. Slattery ML, Edwards SL, Boucher KM, Anderson K, Caan BJ. Lifestyle and colon cancer: an assessment of factors associated with risk. Am J Epidemiol 1999;150(8):869–77. 3. Center MM, Jemal A, Ward E. International trends in colorectal cancer incidence rates. Cancer Epidemiol Biomarkers Prev 2009;18(6): 1688–94. 4. Edwards BK, Ward E, Kohler BA, et al. Annual report to the nation on the status of cancer, 1975–2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010;116(3):544–73. 5. Giovannucci E, Wu K. Cancers of the colon and rectum. In: Cancer epidemiology and prevention, 3rd ed. New York: Oxford University Press, 2006:809–29. 6. U.S. Food and Drug Administration. Statement of general policy or interpretation. Subchapter B—Food and food products, Part 121— Food additives. Fed Register 1973;38:20725–6. 7. Mason JB, Dickstein A, Jacques PF, et al. A temporal association between folic acid fortification and an increase in colorectal cancer rates may be illuminating important biological principles: a hypothesis. Cancer Epidemiol Biomarkers Prev 2007;16(7):1325–9. 8. Osterhues A, Holzgreve W, Michels KB. Shall we put the world on folate? Lancet 2009;374(9694):959–61. 9. McGlynn AP, Wasson GR, O’Reilly SL, et al. Low colonocyte folate is associated with uracil misincorporation and global DNA hypomethylation in human colorectum. J Nutr 2013;143(1):27–33.
S72
Keum and Giovannucci / Am J Prev Med 2014;46(3S1):S65–S72
10. Giovannucci E, Stampfer MJ, Colditz GA, et al. Multivitamin use, folate, and colon cancer in women in the Nurses0 Health Study. Ann Intern Med 1998;129(7):517–24. 11. Lee JE, Willett WC, Fuchs CS, et al. Folate intake and risk of colorectal cancer and adenoma: modification by time. Am J Clin Nutr 2011;93(4): 817–25. 12. Winawer SJ. Natural history of colorectal cancer. Am J Med 1999;106 (1A):3S–6S; discussion 50S–51S. 13. CDC. Folic acid recommendations. 2012 January 13. www.cdc.gov/ ncbddd/folicacid/recommendations.html. 14. Briefel RR, Johnson CL. Secular trends in dietary intake in the U.S. Annu Rev Nutr 2004;24:401–31. 15. Bentley TG, Willett WC, Weinstein MC, Kuntz KM. Population-level changes in folate intake by age, gender, and race/ethnicity after folic acid fortification. Am J Public Health 2006;96(11):2040–7. 16. Wu K, Platz EA, Willett WC, et al. A randomized trial on folic acid supplementation and risk of recurrent colorectal adenoma. Am J Clin Nutr 2009;90(6):1623–31. 17. Gibson TM, Weinstein SJ, Pfeiffer RM, et al. Pre- and postfortification intake of folate and risk of colorectal cancer in a large prospective cohort study in the U.S. Am J Clin Nutr 2011;94(4):1053–62. 18. Schoen RE, Pinsky PF, Weissfeld JL, et al. Colorectal-cancer incidence and mortality with screening flexible sigmoidoscopy. N Engl J Med 2012;366(25):2345–57. 19. Kim DH, Smith-Warner SA, Spiegelman D, et al. Pooled analyses of 13 prospective cohort studies on folate intake and colon cancer. Cancer Causes Control 2010;21(11):1919–30. 20. Stevens VL, McCullough ML, Sun J, Jacobs EJ, Campbell PT, Gapstur SM. High levels of folate from supplements and fortification are not associated with increased risk of colorectal cancer. Gastroenterology 2011;141(1):98–105, 105 e1.
21. Cole BF, Baron JA, Sandler RS, et al. Folic acid for the prevention of colorectal adenomas: a randomized clinical trial. JAMA 2007;297(21):2351–9. 22. Qin X, Cui Y, Shen L, et al. Folic acid supplementation and cancer risk: a meta-analysis of randomized controlled trials. Int J Cancer 2013;133 (5):1033–41. 23. Kalmbach RD, Choumenkovitch SF, Troen AM, D’Agostino R, Jacques PF, Selhub J. Circulating folic acid in plasma: relation to folic acid fortification. Am J Clin Nutr 2008;88(3):763–8. 24. Mason JB. Folate consumption and cancer risk: a confirmation and some reassurance, but we’re not out of the woods quite yet. Am J Clin Nutr 2011;94(4):965–6. 25. Ulrich CM, Potter JD. Folate and cancer—timing is everything. JAMA 2007;297(21):2408–9. 26. Vollset SE, Clarke R, Lewington S, et al. Effects of folic acid supplementation on overall and site-specific cancer incidence during the randomised trials: meta-analyses of data on 50,000 individuals. Lancet 2013;381(9871):1029–36. 27. Lin YW, Wang JL, Chen HM, et al. Folic acid supplementary reduce the incidence of adenocarcinoma in a mouse model of colorectal cancer: microarray gene expression profile. J Exp Clin Cancer Res 2011;30:116. 28. Guessous I, Dash C, Lapin P, Doroshenk M, Smith RA, Klabunde CN. Colorectal cancer screening barriers and facilitators in older persons. Prev Med 2010;50(1–2):3–10. 29. Fenton JJ, Cai Y, Green P, Beckett LA, Franks P, Baldwin LM. Trends in colorectal cancer testing among Medicare subpopulations. Am J Prev Med 2008;35(3):194–202. 30. Gao QY, Chen HM, Chen YX, et al. Folic Acid prevents the initial occurrence of sporadic colorectal adenoma in Chinese older than 50 years of age: a randomized clinical trial. Cancer Prev Res (Phila) 2013;6(7):744–52. 31. CDC. Second national report on biochemical indicators of diet and nutrition in the U.S. population. CDC, 2012. www.cdc.gov/nutritionreport/.
www.ajpmonline.org
