Let’s now put all the harm (NNH) and benefit (NNT) pieces together. With an NNH for multiple spiral whole-body CT scans at about 100—the greatest cancer risk we’ve seen yet from a diagnostic radiographic procedure—and with an NNT that is currently undetermined but not likely to be any lower than mammography’s NNT of 1,000, the value of using spiral whole-body CT scanning to screen for disease seems questionable. (Remember, for NNH values, higher numbers are better because it means you have to expose a lot more people before one is harmed. In contrast, for the NNT, lower numbers are better because it means you don’t need to treat as many people in order to get one person who benefits.) Thus, the NNH and NNT numbers suggest that overall it is more likely to harm, rather than benefit, the screened population. That’s not to say the procedure doesn’t have value for patients that present with symptoms of illness. It’s just that the “worried well” should beware.
WHAT ELSE?
There are many diagnostic radiography procedures out there and we can’t review them all. Nevertheless, the three procedures highlighted here—arm x-ray, mammography screening, and spiral whole-body CT screening—like all other diagnostic radiography procedures, fall into one of two categories of use: (1) finding disease in patients that present with clinical symptoms of disease; or (2) screening for disease in people who have no symptoms.
In the first category, the benefits nearly always far outweigh the risks, as long as the chosen procedure is appropriate for the clinical condition. This is because the penalty for not finding the underlying cause of the disease can be severe—a nonfunctioning arm or death from cancer—and the risks of partial body radiography procedures are quite low.
In the second category, regarding the screening of the worried well, we should take pause. Do the real benefits meet the claims? How does the NNT compare to the NNH? And what are the added perils apart from the radiation? All these things need to be considered in consultation with a knowledgeable physician before making any decisions.
The issues are complicated, particularly if a person has a family history of disease. It could be argued that a positive family history is, in and of itself, a clinical finding that warrants radiographic screening. But everyone—patients as well as doctors—needs to understand that such screenings also have false-negative findings, in addition to the false-positive findings we’ve discussed. In the case of mammography, a false negative is a mammogram report claiming a clean bill of health for a woman who actually does have breast cancer.16 Although the false-negative problem is much smaller than the false-positive problem, and gets much less attention, false negatives do occur. For every one woman saved by participating in a screening program, six women will die despite their participation in screening.17 Women need to understand that participating in a screening program does not guarantee breast cancer survival.
WHAT ABOUT RADIATION THERAPY?
Even with all its nuances, weighing the risks of diagnostic radiography procedures is a cakewalk compared to weighing the risks and benefits of therapeutic radiation. Why is that? Because radiation therapy is primarily used to treat people who already have cancer, and that makes all the difference. Every patient and every cancer is different, and even the same cancer in the same patient can respond differently to radiation therapy over time. Also, radiation is often combined with both surgery and chemotherapy, because cancer is a formidable foe. This further complicates things.
Despite all these complexities in weighing the risks and benefits of radiation therapy, most radiation risk calculations for cancer patients are a moot point anyway. Many clinical and personal issues need to be weighed when considering whether or not to undergo radiation therapy for cancer. Usually, however, the risk of getting a secondary cancer from radiation treatment for the first one is not one of the primary concerns. This is because a cancer produced by radiation today will typically not appear for 10 to 30 years. It makes little sense to forgo treatment of a real cancer today based on a concern about a theoretical cancer that won’t appear for at least 10 years anyway. The risk of dying now is greater than your risk of dying then. Also, for many older people, the latency period of the second cancer is longer than their natural life expectancy anyway. So they most likely will not live long enough to experience the second cancer. Secondary cancers are of more concern for younger cancer patients who have longer life expectancies.
On an optimistic note, we should welcome the day when secondary cancer risks come to be the major concern for radiation therapy patients. Why? Because it will mean that radiation therapy has become so successful at curing patients that we now have significant reason to worry about what will happen to them decades down the road. The good news is that we already seem to be headed in that direction. Currently, there are more than 12 million cancer survivors in the United States, three times as many as there were in 1971, and the number grows by about 2% each year.18 As this population grows, so will the number of secondary cancer diagnoses, and the medical community will find itself a victim of its own cancer therapy successes. Let’s hope this trend continues and the risk of secondary cancers from radiation therapy becomes the main radiation concern for the next edition of this book.
CHAPTER 14
SORRY, WRONG NUMBER: CELL PHONES
The absence of evidence is not evidence of absence.
—Martin Rees
Facts are stubborn, but statistics are more pliable.
—Mark Twain
TROUBLE ON SCOTTOW HILL
Scarborough is a small town of about 20,000 people on the southern coast of Maine. Now a resort area, it was once the site of many a skirmish with the local Abenaki Indians who objected to European settlers squatting on their land. The attraction for the settlers was excellent fishing and farmland, and ample stream flow to power sawmills, so they resisted moving on, even though they were under constant threat of assault.
What the town needed was a means to communicate danger to everyone in the event of an impending attack. Centuries before the advent of telecommunications, they didn’t have many options, but they saw potential on Scottow Hill, one of the few elevated points of land. The settlers built a tower of wood on top of the hill, to be lit in case of an emergency. If any residents saw a bonfire on top of Scottow Hill, they knew to take refuge, for danger was imminent.
Today the Abenaki Indians are no longer a threat, and the bonfire tower is gone, now replaced by a cell phone tower that serves a similar purpose of ensuring vital communication among the residents. It also protects them from danger, but perhaps not quite as well. It is one of three cell phone towers in Scarborough, but service in town is still spotty. Resident Donald Day is concerned about the well-being of his family. He says the spotty coverage is a safety issue when traveling after dark on icy winter roads through the Scarborough Marsh. Day is a supporter of a proposal before the town council to erect another tower to improve cell phone coverage.
But other residents have their own safety concerns. Elisa Boxer is worried about the effect cell phone towers have on the health of residents. At a town council meeting in the summer of 2014, she gave a statement to the commissioners: “When a cell phone tower goes up in a neighborhood, it becomes a sick neighborhood. When it goes up near a school, it becomes a sick school and the teachers and students become part of a cancer cluster.” Boxer went on to claim that radio waves from cell phone towers have been “scientifically linked” to cancer. Resident Suzanne Foley-Ferguson shared Boxer’s fears. She said the cell towers needed to be “kept as far from any residential neighborhood as possible.” Another resident, Karen Tanguay, chimed in that more studies were needed to truly figure out the health dangers from cell phone technology.
The outcome of the meeting was that the town council ducked the controversy and sent the cell tower proposal back to the ordinance committee. “I am glad it’s going back to Ordinance,” Town Councilor William Donovan said. “As much as it’s been worked through at several levels, I think the public needs to know more about it.” If
Day wants to keep his family safe on the roads, he’ll need to drive around the Scarborough Marsh on wintry nights, because there will be no cell phone coverage there any time soon.1
AN EPIDEMIC IN SEARCH OF A DISEASE
In 2011, the International Agency for Research on Cancer (IARC) in Lyon, France—a wing of the World Health Organization (WHO)—announced that its expert committee had ruled that radiofrequency electromagnetic fields (i.e., radio waves) should be classified among “possible carcinogens.” The announcement was based largely on the findings of a very large cohort study and five smaller case-control studies,2 and the world’s news organizations took notice. The cell-phone opponents proclaimed victory, while the cell-phone fans screamed foul. The committee’s ruling was actually a split decision between a majority of scientists who felt the data were strong enough to say it was “possible,” and a minority who felt that the data weren’t strong enough to say even that.
Although cell phone towers afford a looming monument to modern technology and a focal point for community anxiety, the highest doses of radio waves to humans actually come from use of the cell phones themselves, not the towers. This is because despite their weaker energy output, the cell phones are held next to the head. As we know, health risks are driven by dose. So if the radio waves were to cause cancer, the most likely place to find an association would be in the tissues receiving the highest doses. If you can’t show that cell phones cause cancers of the head, it is not likely that you will be able to show that other cancers, in tissues that receive only a tiny fraction of the head’s dose, are caused by radio waves. So most cell phone studies focus on cancers of the head, more specifically brain cancers, since the volume of the brain is quite large and absorbs most of the radio wave energy.
The cohort study was huge, including 420,095 Danish cell phone subscribers between 1982 and 1995.3 This is much larger than the atomic bomb cohort study, which only follows 120,000 survivor and control subjects. The Danish study showed no association between brain cancers and cell phone use.
Cohort studies, as we’ve learned, are the gold standard of epidemiology and are seldom unseated by the less reliable case-control studies. Normally such a large cohort study with negative findings would have pushed the smaller case-control studies to the sidelines, but there were issues with the cohort study’s design. The chief problem was that doses were never actually measured. Rather than using an exposure metric that can be directly related to dose, just having a cell phone subscription was taken as an indication of a person’s exposure to cell phone radio waves. Perhaps the person owned a cell phone but never used it, or perhaps someone who was not a subscriber was using a cell phone owned by another person. The assumptions that the cell phone subscribers were the users and that their doses were proportional to the length of their cell phone subscription were seen to be weaknesses of the study, and this caused it to drop from gold to bronze status in the eyes of some committee members. This dosimetry problem opened the door for this extremely large negative cohort study to be shelved in favor of case-control studies that had better dose information.
Several small case-control studies suggested that there might be an association between gliomas, a specific type of brain cancer, and cell phone use, but they provided no evidence for an association with meningiomas (tumors of the brain’s surface membrane), salivary gland tumors, leukemias, or lymphomas. Unfortunately, most of these case-control studies had their own issues, including the continual concern regarding all case-control studies; that is, there might be hidden biases in the study design. But one large case-control study was considered robust enough to be given serious attention. The INTERPHONE study, the largest cell phone case-control study to date, compared glioma patients (2,708 cases) to healthy individuals (2,972 controls). The study employed hours of cell phones use as a relative measure of radio wave dose to the brain.4 The study found that for those in the top dose decile5 (i.e., top 10%) of cell phone use (>1,640 total hours), there was a statistically significant increase of 40% in the risk of gliomas, compared to people who never used a cell phone. In addition, the elevated risk was found only in patients who had been exposed for at least seven years. Shorter exposures showed no increased risk, which was consistent with the idea that if an association were genuine there should be a lag, or latency period, before tumors would be expected to appear, as happens for all carcinogens. Worrisome findings.
But when the risk at the lower doses was examined the picture looked much fuzzier. Not only was there no elevated risk at any of the lower dose deciles, but for some deciles there was even statistically significant protection from gliomas! So the data did not support a clear dose-response relationship. And the qualitative shift from protection to elevated risk raised concerns that the study’s findings were fallacious.
Despite the problem of bouncing associations between dose and risk from one decile to the next, what if the risk level at the highest decile were reliable? Furthermore, although it was a case-control study and, therefore, potentially subject to various types of biases, it did not appear to be subject to any recall bias. This is because the hours of cell phone use for each individual in the study were objectively assessed from their phone records, rather than by interviewing the patients about their phone use behavior. In the absence of any identifiable biases, the committee felt that it had no alternative but to take the 40% increased risk of glioma at the highest decile of cell phone use at face value and acknowledge that it was “possible” that radio waves cause cancer.
Forty percent sounds bad, but how bad is it, really? Let’s calculate the NNH to give us a sense of the level of risk, using the conservative assumption that the high-dose findings of the INTERPHONE study are real and that every cell phone user has the same alleged 40% risk as those in the highest dose decile. To put it another way, let’s ask the following question: What would a 40% overall increase in risk mean to those 420,095 Danish cell phone subscribers from the cohort study if all of them, not just those with the most exposure, were at 40% increased brain cancer risk from their cell phones?
In any group of 420,095 people, about 2,500 will develop a brain cancer (or other nervous system cancer) at some point during their lifetime.6 If all of these people were lifelong cell phone users, and their cell phone use really does increase their risk of brain cancer by 40%, we would expect to see 3,500 cancer cases instead of 2,500 within this group (i.e., 1,000 more cases). By dividing 420,095 people by 1,000 additional cases, we get an NNH of 420; this means that among 420 lifetime cell phone users, only one person might be expected to develop a brain tumor due to his or her phone use. Still, given the large number of cell phone users in the United States (91% of adults and 78% of teenagers),7 it suggests that brain cancers produced by cell phones could be generating a substantial increase in brain cancers nationwide—if it were, in fact, true that cell phones cause brain cancer.
HABEAS CORPUS
If cell phones are in fact producing all of these cancers, where are the cancer patients? The fact is that the brain cancer incidence in the United States has remained relatively unchanged for 40 years—a period of time over which cell phone use has increased among adults from 0% in 1980 to 91% as of 2013, with the most rapid increase occurring between 1995 and 2005. There are now more mobile devices in the United States than there are people. When cigarette smoking started to become popular and the cigarette market exploded, there was a huge increase in lung cancers shortly afterward and the elevated rates of lung cancer didn’t begin to subside until cigarette smoking started to drop. Why hasn’t this pattern happened with cell phones? We have to ask: “If cell phones are truly killing us with brain cancer, where are the bodies?”
The data suggest that we do not have an increasing trend for glioma incidence in either men or women, for any age group.8 We have the same glioma rates now, after decades of cell phone use, as we did before cell phones were invented. So even if it is theoretically possible that cell phones could cause brain cancer, they don’t seem t
o be doing so in any great numbers.
It could be argued that if only the most highly exposed people were at significant risk of glioma, it might not translate into a noticeable increase in overall glioma rates. That is, if only the top 10% were affected, we might not be able to see an increase across the whole population. Point taken. If that’s the case, however, it still suggests that radio wave-induced glioma is not a significant public health threat for the majority of people, and few (or none) of the glioma cases that occur annually can be attributed to cell phones. In any case, gliomas are not a public health threat that is on the rise, so there appears to be no reason to believe that we are facing a surge of brain cancers caused by cell phones or, for that matter, any other technologies introduced over the last 40 years.
PRELIMINARY HEARING
Some people are concerned that it just may be too soon to tell whether cell phones cause cancer. As we know from studies with ionizing radiation, there is a latency period between exposure and the occurrence of cancer. But latency is more related to cancer type, rather than the type of carcinogen. Ionizing radiation-induced brain tumors began to appear between five and ten years after the atomic bombings in Japan. So we would expect the same latency for cell-phone-induced brain tumors, if they were to occur. We haven’t seen rates increase yet, but perhaps we just haven’t waited long enough … perhaps. However, radio waves are not exclusive to cell phones. They have been used in telecommunications since the time of Marconi, and the early radio-wave workers were exposed to massive doses because Marconi erroneously assumed that the only way to increase transmission distance was to boost power output to extreme levels. As we’ve already learned, workers are usually the first to bear the consequences of exposure to hazardous agents because they typically experience the highest doses for the longest periods of time. The history of radio waves over the last 100 years suggests that those workers are not at increased risk of cancer. In contrast, the increases in cancer among ionizing radiation workers were very evident within a few years after x-rays were discovered. Cell phones emit just a minuscule fraction of radio-wave energy that early radio workers were exposed to, many of whom worked on transmitting towers while the towers were transmitting, with their heads right in the path of the radio waves. So how can cell phones be producing cancer among the general public when we don’t see cancers among the workers? Good question.
Strange Glow Page 33