Mammogram Recalls Not Affected by HT Suspension
Buist DS, Anderson ML, Reed SD, et al. Short-term hormone therapy suspension and mammography recall: a randomized trial. Ann Intern Med 2009;150:752-765.
Because hormone therapy (HT) increases breast density and can result in abnormal mammograms, some clinicians recommend a short-term suspension of HT before mammography. This randomized, controlled trial (RCT) by Buist et al tested whether 1 to 2 months of HT suspension decreases recall rates for further imaging in women ages 45 to 80. A total of 1,704 women from western Washington between 2004 and 2007 who were using HT at their last mammogram (index) and due for their next screening (study) were randomly divided into three groups (by breast density and HT type). The groups were no HT suspension (n = 567), 1-month suspension (n = 570), or 2-month suspension (n = 567) before study mammography. One blinded expert radiologist interpreted all mammograms. The primary outcome was recall rate, and the secondary outcome was change in mammographic breast density (percentage and dense area) between the index and study mammograms.
Recall rates were 11.3% in the no-suspension group (61 of 542 women), 12.3% in the 1-month suspension group (50 of 478 women), and 9.8% in the 2-month suspension group (44 of 451 women). None of the groups showed a decrease in mammography recalls. Decreases in the percentage of breast density were orderly and statistically significant: 0.1% for the no-suspension group; -0.9% for the 1-month suspension group; and -1.5% for the 2-month suspension group. Similar ordered decreases were observed for dense area. Another result was that women in the suspension groups experienced increased menopausal symptoms. The authors stated that limitations to the study were that results can only be generalized to women ages 45 to 80 who have used HT for at least 1 year and will consider short-term suspension (61% of eligible women declined participation) and that only one expert radiologist determined the recalls.
The study's conclusion was that HT suspension was associated with small changes in breast density and did not affect recall rates. There was no evidence to support short-term HT suspension before mammography.
Commentary by Wendy Y. Chen, MD, MPH
Multiple studies have shown that current hormone therapy (HT) use can be associated with increased breast density and higher rates of mammography recall. After discontinuation of HT, breast density and mammography recall rates both return to baseline. Buist et al attempt to address the important clinical question of whether "short-term" suspension of HT would affect mammography recall rates and improve mammography performance. As the authors acknowledge, the "gold standard" trial would have as its primary outcome the sensitivity and specificity of mammography, but such a trial would need to enroll hundreds of thousands of women. Instead, the authors chose surrogate outcomes, including mammography recall rates and mammographic density, to evaluate the effect of HT suspension on mammography performance.
Recall rates were based on the study radiologist's recommendations and included recalls due to technical issues. It should be noted that there were different rates of withdrawal from the assigned intervention groups, with more withdrawals from the suspension groups: 4.2% from the no-suspension group, 15.8% from the 1-month group, and 20.1% from the 2-month group (no P value given for comparison). In addition, adherence also differed across the groups, with lower adherence among the suspension groups: 99.0% in the no-suspension, 92.7% in the 1-month group, and 87.3% in the 2-month group (P < 0.006 comparing 1- to 2-month suspension groups). Only 1,471 women had complete assessments and were evaluable for the primary analysis. In modified intention-to-treat analyses in which only women who did not withdraw from the study prior to mammography were included, there was no difference in recall rates. Modest changes in breast density were seen with HT suspension, driven mainly by changes among women who suspended use of estrogen plus progestogen. As expected, women within the suspension groups reported more menopausal symptoms.
The use of recall rates as a surrogate outcome was reasonable, since most recalls are due to false-positive screens, an important clinical issue that leads to increased medical costs and patient anxiety. However, there are several important limitations to the study. Although statistical power was estimated by the authors at 85% to detect a change in recall rates from 13% to 8%, this would be a very striking decrease and one that would not likely be observed with such a brief intervention, which means that realistically, the study was a bit underpowered. In addition, it is not clear how generalizable the findings would be to routine clinical care because the recall rates were based on a single study radiologist reviewing all films and included recalls due to technical reasons (eg, insufficient tissue in one or more views in order to compare to previous views). Data were not provided on the rates of concordance between the recommendations of the study radiologist and the radiologists included as part of routine clinical care. Finally, as detailed previously, withdrawal and adherence rates were an issue and were strongly related to the treatment assignment. This, along with use of modified intention-to-treat analyses (rather than a true intention-to-treat analysis), interferes with the principles of randomization and may introduce bias.
In conclusion, this randomized trial did not observe any difference in mammography recall rates (and, by implication, mammography performance) with brief suspension of HT use, but this study had several weaknesses including limited statistical power and issues regarding completeness of randomization. However, all women and providers remember that long-term HT using combination estrogen and progestogen are no longer routinely recommended for women after the findings of the Women's Health Initiative in which overall health risks exceeded overall health benefits.
From the NAMS First to Know e-newsletter released September 24, 2009
For more, please visit http://www.menopause.org/news.aspx
Summary
Buist DS, Anderson ML, Reed SD, et al. Short-term hormone therapy suspension and mammography recall: a randomized trial. Ann Intern Med 2009;150:752-765.
Because hormone therapy (HT) increases breast density and can result in abnormal mammograms, some clinicians recommend a short-term suspension of HT before mammography. This randomized, controlled trial (RCT) by Buist et al tested whether 1 to 2 months of HT suspension decreases recall rates for further imaging in women ages 45 to 80. A total of 1,704 women from western Washington between 2004 and 2007 who were using HT at their last mammogram (index) and due for their next screening (study) were randomly divided into three groups (by breast density and HT type). The groups were no HT suspension (n = 567), 1-month suspension (n = 570), or 2-month suspension (n = 567) before study mammography. One blinded expert radiologist interpreted all mammograms. The primary outcome was recall rate, and the secondary outcome was change in mammographic breast density (percentage and dense area) between the index and study mammograms.
Recall rates were 11.3% in the no-suspension group (61 of 542 women), 12.3% in the 1-month suspension group (50 of 478 women), and 9.8% in the 2-month suspension group (44 of 451 women). None of the groups showed a decrease in mammography recalls. Decreases in the percentage of breast density were orderly and statistically significant: 0.1% for the no-suspension group; -0.9% for the 1-month suspension group; and -1.5% for the 2-month suspension group. Similar ordered decreases were observed for dense area. Another result was that women in the suspension groups experienced increased menopausal symptoms. The authors stated that limitations to the study were that results can only be generalized to women ages 45 to 80 who have used HT for at least 1 year and will consider short-term suspension (61% of eligible women declined participation) and that only one expert radiologist determined the recalls.
The study's conclusion was that HT suspension was associated with small changes in breast density and did not affect recall rates. There was no evidence to support short-term HT suspension before mammography.
Commentary by Wendy Y. Chen, MD, MPH
Multiple studies have shown that current hormone therapy (HT) use can be associated with increased breast density and higher rates of mammography recall. After discontinuation of HT, breast density and mammography recall rates both return to baseline. Buist et al attempt to address the important clinical question of whether "short-term" suspension of HT would affect mammography recall rates and improve mammography performance. As the authors acknowledge, the "gold standard" trial would have as its primary outcome the sensitivity and specificity of mammography, but such a trial would need to enroll hundreds of thousands of women. Instead, the authors chose surrogate outcomes, including mammography recall rates and mammographic density, to evaluate the effect of HT suspension on mammography performance.
Recall rates were based on the study radiologist's recommendations and included recalls due to technical issues. It should be noted that there were different rates of withdrawal from the assigned intervention groups, with more withdrawals from the suspension groups: 4.2% from the no-suspension group, 15.8% from the 1-month group, and 20.1% from the 2-month group (no P value given for comparison). In addition, adherence also differed across the groups, with lower adherence among the suspension groups: 99.0% in the no-suspension, 92.7% in the 1-month group, and 87.3% in the 2-month group (P < 0.006 comparing 1- to 2-month suspension groups). Only 1,471 women had complete assessments and were evaluable for the primary analysis. In modified intention-to-treat analyses in which only women who did not withdraw from the study prior to mammography were included, there was no difference in recall rates. Modest changes in breast density were seen with HT suspension, driven mainly by changes among women who suspended use of estrogen plus progestogen. As expected, women within the suspension groups reported more menopausal symptoms.
The use of recall rates as a surrogate outcome was reasonable, since most recalls are due to false-positive screens, an important clinical issue that leads to increased medical costs and patient anxiety. However, there are several important limitations to the study. Although statistical power was estimated by the authors at 85% to detect a change in recall rates from 13% to 8%, this would be a very striking decrease and one that would not likely be observed with such a brief intervention, which means that realistically, the study was a bit underpowered. In addition, it is not clear how generalizable the findings would be to routine clinical care because the recall rates were based on a single study radiologist reviewing all films and included recalls due to technical reasons (eg, insufficient tissue in one or more views in order to compare to previous views). Data were not provided on the rates of concordance between the recommendations of the study radiologist and the radiologists included as part of routine clinical care. Finally, as detailed previously, withdrawal and adherence rates were an issue and were strongly related to the treatment assignment. This, along with use of modified intention-to-treat analyses (rather than a true intention-to-treat analysis), interferes with the principles of randomization and may introduce bias.
In conclusion, this randomized trial did not observe any difference in mammography recall rates (and, by implication, mammography performance) with brief suspension of HT use, but this study had several weaknesses including limited statistical power and issues regarding completeness of randomization. However, all women and providers remember that long-term HT using combination estrogen and progestogen are no longer routinely recommended for women after the findings of the Women's Health Initiative in which overall health risks exceeded overall health benefits.
From the NAMS First to Know e-newsletter released September 24, 2009
For more, please visit http://www.menopause.org/news.aspx
Source...