Respectful Insolence

I was perusing some articles that had accumulated while I was away, looking for ones that I wouldn’t want to have missed and also looking for blog fodder (sometimes my day job and my blogging job actually mesh quite well, at least when it comes to discussing biomedical studies), and then I found an article that I had to discuss. Somehow I had missed it in the week leading up to my departure; how that happened I don’t know, but it’s time to make up for it. In any case, when I saw this article over the weekend, I knew I’d better comment on it, because, mark my words, it will soon be showing up on all sorts of altie websites as “proof” that chemotherapy does more harm than good and that doctors are out only to poison cancer patients, all for the better profit of big pharma. Never mind that it shows nothing of the sort; it is still an important and, to those of who take care of breast cancer patients, a somewhat disturbing study that could well have an impact on our recommendations regarding who should and should not receive postoperative adjuvant chemotherapy for breast cancer.

Here’s the article’s citation:

Hassett MJ, J O’Malley, JR Pakes, JP Newhouse, and CC Earle. Frequency and Cost of Chemotherapy-Related Serious Adverse Effects in a Population Sample of Women With Breast Cancer. J Natl Cancer Inst 98:1108-1117 (2006).

And here’s how the paper is being represented in the media:


Chemo harms more breast cancer patients

Uh-oh. The article itself is a fairly balanced assessment of the study, but the headline made me cringe. Here’s how the AP described the study:

WASHINGTON – Younger breast cancer patients seem to suffer more serious side effects from chemotherapy than previously thought. Roughly one in six of those women wind up at the emergency room or hospitalized because of such side effects as infection, low blood counts, dehydration or nausea, researchers reported Tuesday.

Some of the side effects occurred at rates three to four times higher than earlier research had predicted.

Tuesday’s study marks the first attempt to assess the real-world risks of chemotherapy for some 35,000 breast cancer patients under age 64 who get the drugs each year.

Most side-effect information comes from clinical trials of medications that can underestimate toxicity. Those trials are designed to prove if the drugs fight cancer and therefore should be sold, and they tend to enroll only the best candidates instead of women who might be particularly sensitive to side effects.

Adding to that conundrum: Many breast cancer patients don’t need chemotherapy in the first place; surgery, radiation and hormone treatment are enough. But doctors don’t always have an easy way to tell who would benefit from chemo on top of all that.

And for women in the to-treat-or-not gray zone, age sometimes is the deciding factor — because those under 64 are thought to tolerate chemotherapy better than older women.

“We don’t believe our study is saying that chemotherapy is not helpful,” stressed Dr. Michael Hassett of Boston’s Dana-Farber Cancer Institute, who led the research, published in Tuesday’s Journal of the National Cancer Institute.

But, “we’ve been struggling as a professional community to understand which women benefit from chemotherapy,” he added.

Here’s the abstract of the article:

Background: The number, nature, and costs of serious adverse effects experienced by younger women receiving chemotherapy for breast cancer outside of clinical trials are unknown. Methods: From a database of medical claims made by individuals with employer-provided health insurance between January 1998 and December 2002, we identified 12 239 women 63 years of age or younger with newly diagnosed breast cancer, of whom 4075 received chemotherapy during the 12 months after the initial breast cancer diagnosis and 8164 did not. Diagnostic codes for eight chemotherapy-related adverse effects were identified. Total hospitalizations for all causes, hospitalizations or emergency room visits for adverse effects that are typically related to chemotherapy, and health care expenditures were compared between the two groups of women. All statistical tests were two-sided. Results: Women who received chemotherapy were more likely than those who did not to be hospitalized or to visit the emergency room for all causes (61% versus 42%; mean difference = 19%, 95% confidence interval [CI] = 16.7% to 21.3%, P<.001) and for chemotherapy-related serious adverse effects (16% versus 5%, mean difference = 11%, 95% CI = 9.6% to 12.4%, P<.001). The percentages of chemotherapy recipients who were hospitalized or visited the emergency room during the year after their breast cancer diagnosis were 8.4% for fever or infection; 5.5% for neutropenia or thrombocytopenia; 2.5% for dehydration or electrolyte disorders; 2.4% for nausea, emesis, or diarrhea; 2.2% for anemia; 2% for constitutional symptoms; 1.2% for deep venous thrombosis or pulmonary embolus; and 0.9% for malnutrition. Chemotherapy recipients incurred large incremental expenditures for chemotherapy-related serious adverse effects ($1,271 per person per year) and ambulatory encounters ($17,617 per person per year). Conclusions: Chemotherapy-related serious adverse effects among younger, commercially insured women with breast cancer may be more common than reported by large clinical trials and lead to more patient suffering and health care expenditures than previously estimated.

This study is important because it attempts to quantify the the rate at which adverse reactions to chemotherapy occur and result in visits to the emergency room and/or hospitalization in the “real world,” rather in the somewhat rarified environment of clinical trials. It’s also important because it focused on chemotherapy complications in younger women with breast cancer. Clinical trials are indeed critical to determine whether a therapy is (1) efficacious and (2) safe. It is the balance between (1) and (2) that determines the worth of a treatment. Does it produce a large amount of benefit with a small rate and severity of side effects or does it produce only a small benefit with a high rate and severity of side effects? The answer to these questions determine whether it is advisable for the new therapy to be applied widely to the general population with the illness that the drug or treatment being tested is designed to treat. It also allows physicians to give accurate estimates of the chances of side effects to patients, helping them to decide if after their surgery they want to go through chemotherapy for the known benefit or not. What is not generally appreciated by the public is that for early stage breast cancer (tumors less than 2 cm in diameter, no positive lymph nodes) the benefits of chemotherapy are actually rather modest, usually less than a 4% absolute improvement in the chance of surviving five years or more. Even so, the vast majority of women, when informed of this modest benefit, will still usually opt for chemotherapy, although we tend to be much less enthusiastic about recommending chemotherapy after surgery for women over 70 or women with health problems that make it less likely that they wiill be able to tolerate chemotherapy.

Basically, what the investigators did is to examine a database of medical claims made by women with health insurance, looking for women under 64 with a new diagnosis of breast cancer. They then looked at diagnostic codes for eight chemotherapy-related adverse events during the first year after diagnosis, looking at total hospitalizations, hospitalizations for chemotherapy-related adverse events, and total health care expenditures. They then compared the rates of these events between women who received chemotherapy and those who did not. As you might imagine for a study of this type, the most difficult and problematic aspect is matching the two cohorts of patients. For example, women with smaller tumors and lower stage breast cancer would by the standard of care be less likely to receive chemotherapy. Patients with axiillary lymph nodes negative for tumor, even if they did get chemotherapy, would undergo a regimen less intensive and prolonged than patients who have lymph node metastases from their tumors. You might well anticipate that patients with more advanced disease might be more prone to complications regardless of whether they received chemotherapy or not, and you might also imagine that those receiving more intensive chemotherapy would be more likely to have complications than those receiving less intensive chemotherapy. Consequently, trying to match the groups, so that the women not receiving chemotherapy were equivalent to those receiving chemotherapy, is of utmost importance.

I have to confess that I don’t entirely understand the statistics behind how this matching was done, but I do understand the basic concept. The investigators used a propensity score-based matching system. Essentially, for each patient a logistic regression model was created predicting the likelihood of that patient receiving chemotherapy based on age, health plan, region of residence, comorbidity scores, and tumor factors that predict survival. Then, each chemotherapy recipient was paired with a nonchemotharpy patient randomly selected from the subset of patients not receiving chemotherapy whose prediction of receiving chemotherapy was within a small amount of the chemotherapy recipient. This was done to assure that chemotherapy recipients were similar to those not receiving chemotherapy. Even with this matching, unfortunately, there were stiill some small imbalances; for example, chemotherapy recipients tended to be younger than nonrecipients, more likely to have metastatic disease, and more likely to be working full time. The two groups were then examined for the eight parameters associated with chemotherapy side effects: infections/fever; neutropenia (low white blood cell count) or thrombocytopenia (low platelet count); anemia; nausea, vomiting, or diarrhea; dehydration and/or electrolyte abnormality; malnutrition; constitutional symptoms related to treatment; deep venous thrombosis and/or pulmonary embolus. Several other conditions that are rarely associated with chemotherapy were also examined.

Not surprisingly, women receiving chemotherapy were much more likely to suffer one or several of the eight chemotherapy-related effects. Reassuringly, other than headaches, there was no increase in non-chemotherapy-related events in patients receiving chemotherapy. Further, it was found that patients receiving chemotherapy were 19% more likely to make a trip to the emergency room in the first year than those who didn’t. That in itself is not surprising (what would have been more surprising is if no difference was found), but what was surprising was the absolute rate of chemotherapy-related complications, 16%, a rate at least two or three times that predicted by clinical trials. Moreover, chemotherapy resulted in an additional expenditure of $1,271 per person per year for adverse events and $17, 617 for ambulatory encounters.

I’m sure that it won’t be long before this paper makes the rounds on altie websites as “evidence” that chemotherapy is useless. It means nothing of the sort. What it does mean, if it is correct (remember, the fact that it only looks at patients with employer-provided health insurance and it is still not certain that the method used to match the two cohorts is sufficently robust), is that in real world conditions as practiced in the community chemotherapy has a higher rate of complications than it does in the highly regulated world of the clinical trial, in which patients are observed more closely than they usually are in routine clinical practice. One problem with clinical trials is that they may not detect uncommon serious adverse effects or may detect them but generate inaccurate risk estimates. These effects may not become apparent until the regimen is given to many more patients in general practice. Moreover, these complications impose a significant additional health care cost per patient.

Over the last couple of decades, there has been a tendency to expand the indications for chemotherapy.
For example, it used to be that we only recommended chemotherapy for node-negative patients; now many node-negative patients receive it as well, because clinical trials have shown a survival benefit. suported by clinical trials and lulled by data that shows a relatively low rate of complications, we’ve been recommending chemotherapy for older and older patients. What this study might be telling us, even though it was confined to patients under 64) is that maybe the risk-benefit ratio for chemotherapy isn’t quite as favorable as we may have thought for patients with early stage, low risk disease. Although studies have shown that women will often accept chemotherapy for a survival advantage as low as 1-2%, they might not be quite as willing to do so if the risk of significant toxicities is higher, making the risk-benefit ratio less favorable. In such circumstances, we may start to recommend forgoing chemotherapy to older patients and patients with more significant comorbidities. Patients with higher risk disease, where the absolute survival benefit for chemotherapy is considerably higher would still be advised to undergo adjuvant therapy. Also, as much as we don’t like to talk about it, insurance companies and third party payors may be less willing to pay for a therapy with a modest benefit if the the cost due to complications from the treatment is too high.

The problem with chemotherapy given postoperatively in the adjuvant setting is that it’s a blunt instrument. There is no doubt from the data that it measurably improves survival when given after definitive surgery for breast cancer. However, the vast majority of women who receive it probably don’t benefit; they would have been cured just by surgery. Unfortunately, we don’t yet have good methods of determining who will and will not benefit from adjuvant chemotherapy. Genomic medicine gives us hope of producing better instruments to predict who is likely and who is unlikely to benefit; indeed, there are already several microarray (gene chip) assays that appear to identify patients at a high risk of relapse (who are likely to benefit from adjuvant chemotherapy) and distinguish them from patients at a low risk of relapse (whom chemotherapy is highly unlikely to benefit). Certainly such tests, if validated, will help spare many women chemotherapy that is unlikely to benefit them. However, ultimately, what is more likely to produce increased survival with much less toxicity is therapy that is molecularly targeted to the the specific genetic defects in a patient’s cancer. Indeed, that is one of the hottest areas of research in oncology, to develop just such targeted therapies, and for the first time we actually have both the knowledge of such specific defects and the technology to design targeted therapies to exploit them, making this the start of a potential golden age in the development of efficacious, nontoxic therapies.

Comments

  1. #1 R
    September 5, 2006

    What does ‘ambulatory encounters’ mean?

  2. #2 Dawn
    September 5, 2006

    Orac…you have confused me. You say “For example, it used to be that we only recommended chemotherapy for node-NEGATIVE (caps mine) patients; now many node-NEGATIVE (caps mine)patients receive it as well, because clinical trials have shown a survival benefit.” Is one of them supposed to be “node-positive”? If so, which one?

    I hope the alties don’t twist the research results too much!

  3. #3 Samantha Vimes
    September 5, 2006

    Another question this brings to mind is: how does overall health play into the decision to have chemo? For instance, if a person has had serious illnesses before, would that make it more likely she would opt for chemo than if she was the picture of health before discovering the cancer? It seems to me that most people I know become more cautious if they have chronic health problems, and therefore more likely to opt for vigorous treatment– in which case, it might not be the chemotherapy alone accounting for the difference in hospitalization rates.

  4. #4 anonimouse
    September 5, 2006

    R,

    Ambulatory encounters are those that take place outside of the traditional hospital-based, inpatient setting. In other words, if you go to your local health clinic or to your family doctor’s office that’s an ambulatory encounter.

    Orac can correct that if I’m wrong. (or some other doctor can) I’m not a physician, I just work with a few soon-to-be ones.

  5. #5 Sid Schwab
    September 5, 2006

    I posted about this subject (not the paper — the “blunt instrument” aspect of cancer treatment. I referred to it as “Neanderthal,” saying what Orac has said in a different way: we don’t have a way of either knowing exactly who will benefit from treatment, or of giving treatments that affect only the cancer cells. So we do the best we can, which is to base recommendations on good data, which apply well to large groups, but make the recommendation for a given individual, basically, a matter of odds-playing. Given how far we’ve come toward the ideal, however, there’s every reason to predict that beyond the mists there will be major improvements.