June 2012
Feature Story

Anne Paxton

Diagnosing pregnancy with a human chorionic gonadotropin (hCG) immunoassay today is a walk in the park compared with the procedures used a couple of generations ago. But recent research suggests that in the nation’s hospitals, where millions of the tests are ordered each year on urine and serum specimens, some of the most common assumptions about if, when, and how hCG testing should be performed may be incorrect.

“Everyone thinks that hCG testing is very simple, but it’s actually more complex,” says David Grenache, PhD, associate professor of pathology at the University of Utah and medical director of special chemistry at ARUP Laboratories. “It’s a good test, but it’s not a perfect test.”

Many physicians mistakenly believe the urine hCG tests are just as good as the serum tests, and the realities and dangers of false-negatives and false-positives are often played down or ignored. Overuse of hCG testing can also create problems, he points out. “There is misinformation and sometimes a poor understanding of the strengths and limitations of hCG tests.”

In fact, says Ann Gronowski, PhD, professor of pathology and immunology at Washington University School of Medicine in St. Louis, she started doing research on hCG testing because physicians would occasionally call the clinical laboratory and ask how they should interpret inconsistent hCG results. “Many of our studies have come out of questions from physicians about particular patients where they haven’t been able to figure out what to do with them.”

While hCG tests are approved by the Food and Drug Administration only for pregnancy diagnosis, the tests have other clinical uses, Dr. Grenache notes. “There are three others besides its use in diagnosing intrauterine pregnancy. In a workup of suspected ectopic pregnancy, serial hCG testing is used, usually in combination with transvaginal ultrasound. It also has clinical utility as a tumor marker,” particularly for germ cell and gestational trophoblastic tumors. “It can help monitor the response to treatment of patients who have those tumors,” he says. It’s also used as part of a panel of tests in maternal serum screening for fetal aneuploidies, such as fetal trisomy 21.

Several factors can contribute to false-negative or false-positive hCG urine results, Dr. Gronowski notes. Specimen mixup is the usual cause of error, but problems more specifically linked to urine hCG tests include incorrect reading time, insufficient or dilute urine, biochemical pregnancy, or the high-dose “hook effect.” Some of the more general reasons for erroneous reporting, researchers say, are the complexity of the hCG molecule, confusion of nomenclature on the various forms of hCG, lack of awareness by some laboratory personnel of distinctions between forms of hCG, lack of clarity and uniformity in manufacturers’ reagent labeling, and lack of information in product inserts on the specificity of each method to various forms of hCG.

In Dr. Gronowski’s view, the urine qualitative tests are never the preferred tests. “First of all, the urine itself is affected by fluid intake and volume status, and at least in early pregnancy when women have low hCG concentrations, you could drink enough liquid that you’d get a negative. In addition, the qualitative devices are generally set at a little higher cutoff, around 25 mIU/mL, whereas serum, which is not affected by volume status, has a cutoff nearer to 5 mIU/mL.”

The high-dose hook effect is a well-known phenomenon in laboratory medicine, Dr. Gronowski says. It occurs when increasing concentrations of analyte lead to saturation of both the capture antibody and detection antibody, preventing formation of a complete antibody-antigen-antibody sandwich and leading to possibly false-negative results. She co-authored a study focusing on what she calls a “variant hook effect” that affects qualitative urine hCG tests. “The problem is an excess of one form of hCG. Part of the problem is that hCG comes in many different isoforms, and after five weeks of pregnancy, you get much more of one variant form called the β core fragment than you get of the hCG intact molecule. We demonstrated that if you add increasing concentrations of β core fragment to a urine sample that contains intact hCG, there is a dose-dependent inhibition of signal. This is similar to the standard hook effect, except that in many cases the devices don’t recognize β core fragment alone. In other words, the β core fragment only binds one of the two antibodies used in the assay, but it still has the same net effect and causes a false-negative result.”

Her research, reported in an article titled “False-negative results in point-of-care qualitative human chorionic gonadotropin devices due to excess hCG β core fragment” (Clin Chem. 2009;55:1389–1394), has shown that the variant hook effect is a common phenomenon in urine hCG tests. “If manufacturers were able to address this problem,” she says, “it could significantly reduce the number of false-negatives.” But the phenomenon highlights the question of whether what the devices are designed to do fits the hospital’s intended use. “The variant hook effect occurs in women who are more than five weeks pregnant, but the devices were designed originally for supposedly early detection of pregnancy. In the hospital setting, we all have a protocol that any woman who comes into the emergency department or for chemotherapy or a transplant, regardless of age, gets a pregnancy test. That’s fine—except if a device was developed to only measure intact hCG, it has a problem when there’s too much of the variant hCG from five weeks’ gestation and later.”

In 2009, Dr. Gronowski and other authors of the article suggested that laboratories screen devices to identify those in which hCG β core fragment causes the least negative interference in women at five to eight weeks’ gestation. “We’ve picked the device we feel is the best based on our screening. But in addition, when we have to select a new lot, we take some urine that we know is high in β core fragment and we test it on a bunch of different lot numbers, and select the lot that seems to have the least effect.”

Mark A. Cervinski, PhD, director of clinical chemistry and assistant professor of pathology, Dartmouth Medical School, a coauthor of the same study, notes that a follow-up study (reported in Clin Chem. 2009; 55:1885–1886) found that switching devices did not make a difference. “They discovered that by changing the lot numbers on the device, effectively they saw the same problem happen again. So if people think they should switch to a device where false-negative pregnancy results don’t occur, unfortunately that device doesn’t exist.” In the hospital, he says, the simple solution is to order serum quantitative hCG tests. “Then this issue of interference by isoforms becomes irrelevant.”

Another problem of hCG testing is false-positive test results in postmenopausal women, Dr. Gronowski says. “They’re not really a false-positive because postmenopausal women begin to make hCG from the pituitary gland. So the test is correctly identifying hCG; it’s just that it is not hCG of pregnancy. This can delay medical procedures. For instance, a woman with slightly elevated hCG who is in the hospital for a hip transplant creates a clinical dilemma. What should you do? First of all, a postmenopausal woman should not have had an hCG test in the first place. That’s a test utilization issue that people need to keep in mind when they order a qualitative hCG test to rule out pregnancy.”

Since hCG is one of the most common tests done in the laboratory, she points out, even a small percentage of false-negatives or false-positives will affect a large number of people. “It’s probably a couple of times a week that we get a call reporting a negative qualitative and positive quantitative hCG, and I think our physicians have learned to ignore it, which is not a good practice if you’re ordering a test, because ultimately you’re going to believe what you feel like believing.”

These kinds of issues have fueled her belief that urine hCG tests may not be necessary at all in the hospital. “Historically, we didn’t have any kind of quick hCG.” But then in the 1960s, “all of a sudden we got radioimmunoassay for hCG and that was revolutionary. Then really rapid tests were developed that can be done in 20 minutes. Today, we have a great quantitative serum hCG in the lab versus a not-as-good qualitative urine test in the ER and other point-of-care settings. We haven’t stopped to say, hey, maybe we shouldn’t really be doing this urine qualitative anymore because we’ve got this great serum test and it provides much better results.”

There appears to be a growing consensus among many experts that one hCG test—qualitative serum hCG—could safely be jettisoned. Dr. Grenache, who coauthored a 2012 article considering which of several lab tests should be obsolete, focused on the qualitative serum pregnancy test. Use of serum as a sample to do qualitative hCG varies by institution. “From talking to colleagues I’ve heard some don’t really do many at all. But others are like us, and we do quite a few.”

The hypothesis of the study (Should the qualitative serum pregnancy test be considered obsolete? Furtado LV, et al. Am J Clin Pathol. 2012;137:194–202) was that the qualitative serum hCG tests could be abandoned. “The reason we thought so was that the qualitative test cannot be performed at the point of care, most labs that offer a qualitative test also offer a quantitative test, and the quantitative test is generally believed to be more sensitive.” What he and his coauthors wanted to find out was whether a laboratory could abandon the qualitative serum test, offer only a quantitative serum test, and still meet physicians’ turnaround time expectations.

This was the first study to compare the TAT of qualitative and quantitative serum hCG tests and to assess physicians’ perceptions of their clinical performance. In a survey that was part of the study, 49 percent of physicians who responded preferred qualitative over quantitative serum tests for determining pregnancy status. But while qualitative serum TAT is generally faster than quantitative, Dr. Grenache says, “it turns out in our study there were no differences in total TAT when transport time and sample processing time were factored in.”

The authors conclude that qualitative serum hCG results could be replaced by quantitative results, though there is no clear advantage to doing so, because qualitative serum hCG tests have detection limits and performance characteristics similar to those of quantitative tests and are slightly less expensive to perform.

Some labs, Dr. Grenache notes, have opted to continue offering the qualitative serum test but perform a quantitative test when it is ordered. “So if an order for a qualitative hCG comes to the laboratory, you put it on the instrument in the lab, and if the result is less than 5 mIU/mL it’s reported as negative; if the result is greater than 25 mIU/mL, it’s just reported as positive. Results between those two cutoffs are reported as indeterminate. It’s a viable option if labs think they still need to offer that type of service.” But he advises that labs considering this step take a look at their own TATs: “We found there was essentially no difference here, but there might be a difference in other organizations.”

Dr. Cervinski, too, is unsure of the usefulness of a qualitative serum test, and he notes that, of late, this has been a topic of conversation on several listservs. “I’ve never worked in a laboratory where we’ve done it. We’ve always performed quantitative testing. If you have to give serum to the lab, you have to draw the sample and spin it down, so the amount of time it takes to do the assays, whether quantitative or qualitative, would be essentially the same. And a quantitative number can be much more useful if you’re trying to determine whether a patient is, for instance, having an ectopic pregnancy. For the same amount of time and effort, you get much more information with a quantitative test.”

However, Dr. Cervinski points out, analytic precision must continue to be a concern. “Our positivity cutoff for pregnancy tests is 5 mIU/mL; that’s the threshold we’ve defined in the community of laboratorians for saying a result is consistent with pregnancy. But there is a small percentage of non-pregnant women of reproductive age who will have detectible hCG right at or near that threshold. So how we define what is positive could in fact lead to some confusion by providers.”

To further reduce false-positives, Dr. Cervinski would like to see manufacturers develop a urine hCG device not subject to the interference from the variant isoforms. “Perhaps we need two separate kinds of urine hCG devices. We may need a very sensitive device to detect hCG at variable concentrations in early pregnancy, and it may also be beneficial to have a POC device that is not so sensitive to these variant isoforms.”

Perhaps a more important issue, from Dr. Grenache’s perspective, is the over-reliance on urine hCG tests to rule out pregnancy. “I’ve seen hCG tests ordered on women in their 90s, women with hysterectomies. The idea is ‘we’ll do the test on everyone; that way we won’t miss anything.’ But in some situations you do get a positive result in a woman in whom it’s unexpected, and it always causes clinical confusion and can delay necessary therapies,” he says.

At one institution where Dr. Grenache worked, a false-negative result was reported on a young woman who came to the ER with nonspecific abdominal pain. “We did a urine hCG at the point of care and got a negative result and the patient was eventually discharged. But she then presented to a different hospital the next day with a ruptured ectopic pregnancy, a life-threatening condition.” After the ER called the lab to complain that the urine test was lousy and had missed this ectopic pregnancy, “We had to educate them about the strengths and limitations of using urine tests. We said, if you’re working up a patient with suspected ectopic pregnancy, you don’t want to rely on a urine test. They didn’t appreciate the fact that there are many reasons why you might have a negative urine test result.”

“It’s convenient to do these tests at the point of care rather than rely on the serum test, which I think is a better test,” he says. “But in fact, you could argue that urine testing should be abandoned as well. I always tell clinicians if you need to rule in or out the presence of hCG with as much assuredness as a laboratory test can provide, then you really should do the quantitative serum test.” That’s hard to put in practice, though, because there’s no way to get the fast TAT of a urine test with serum. “What we really need are point-of-care blood tests for hCG. That would help a lot, but I don’t know of any company making them.” Some quantitative rapid analysis instruments are in use abroad but have not been cleared for use in the U.S., he notes.

Dr. Gronowski would like to see a good study comparing the turnaround time and false-positive/false-negative results of hCG testing in a big center with and without the availability of a POC urine hCG test. “It’s a difficult study, because you would need to perform a serum parallel test for every single patient that had a urine test done. But it’s very interesting—a few years ago I participated in the NACB LMPG on point-of-care testing. A sub-study examined reproductive testing, and the question we asked was: ‘Does the use of urine hCG POCT as an aid in the diagnosis of pregnancy improve outcomes—that is, reduce clinic visits or reduce length of stay in the emergency department or reduce number of contraindicated drugs or therapies—compared to serum core laboratory hCG?’ Remarkably, there were no studies that had examined this.”

For serum testing, Dr. Cervinski emphasizes, the need is for harmonization. “Right now, if one sample were to go to two different labs using two different manufacturers of reagents, you could get two different results because the hCG assay is not well harmonized. For pregnancy, that may not be much of an issue because a positive is generally a positive. However, if hCG were being used as a tumor marker and a patient were to get lab services at two different places, that might make results appear erroneously to be trending up or trending down.”

To effectively reduce the number of erroneous urine hCG results, hospitals may need to form a multidisciplinary committee to reconsider their point-of-care hCG testing protocols, Dr. Cervinski adds. “It really does seem somewhat excessive to automatically test all women patients coming into the hospital. But it’s not enough to make a change at the administrative level, because staff will need training to help understand the new policy.” That’s the real challenge in the POC field, he says: “We’re not talking about a few dozen people in the laboratory; we’re dealing with an extremely large population whose experience in laboratory testing is quite variable.”