Measuring serum hormone concentrations in postmenopausal women at high risk for developing breast cancer may help identify those most likely to benefit from preventive treatment with an aromatase inhibitor, according to findings from a case-control study using data from a large breast cancer prevention trial.

In the randomized, placebo-controlled IBIS-II prevention trial of 3864 women aged 40-70 years at increased risk for developing breast cancer, treatment with the aromatase inhibitor anastrozole was associated with a 49% reduction in breast cancer incidence. At median follow-up of 131 months, breast cancer occurred in 85 (4.4%) versus 165 (8.5%) of patients in the anastrozole and placebo arms, respectively.

A preplanned case-control study involving 212 participants from the anastrozole group (72 cases and 140 controls) and 416 from the placebo group (142 cases and 274 controls), showed a significant trend toward increasing breast cancer risk with increasing estradiol-to-sex hormone binding globulin (SHBG) ratio in the placebo group, but not in the anastrozole group (trend per quartile, 1.25 vs 1.06), reported Jack Cuzick, PhD, of the Wolfson Institute of Population Health, Queen Mary University of London, UK, and colleagues.

A weaker but still significant effect was observed for the testosterone-SHBG ratio in the placebo group (trend, 1.21), but again, no such effect was seen in the anastrozole group (trend, 1.18).

A relative benefit was seen for anastrozole in estradiol concentration quartiles 2, 3, and 4 (relative risk [RR], 0.55, 0.54, and 0.56, respectively), but not in quartile 1.

The findings were published online December 6 in The Lancet Oncology.

Study participants were recruited from 153 breast cancer treatment centers across 18 countries and randomized in a 1:1 ratio to receive 1 mg of oral anastrozole daily or placebo. For the case-control analysis, the investigators looked at the effects of baseline estradiol to SHBG ratio on the development of all breast cancers, including ductal carcinoma in situ. They also assessed the relative benefit of anastrozole versus placebo.

Case patients were those diagnosed with breast cancer after trial entry through data cutoff on October 22, 2019, and who had not used hormone replacement therapy within 3 months of trial entry or during the trial. Controls were participants without breast cancer who were randomly selected and matched according to treatment group, age, and follow-up time.

“Although the association between estradiol and breast cancer risk is well established, less is known about whether the concentrations of these hormones have an effect on the efficacy of preventive therapy with selective estrogen receptor modulators or aromatase inhibitors in women at increased risk of developing breast cancer,” the investigators noted, explaining that in the current analysis, they “tested the hypothesis that, for women with a low estradiol–SHBG ratio, anastrozole would provide little or no reduction in the risk of breast cancer.”

The results from the placebo group “confirm the increasing risk of breast cancer associated with higher estradiol and testosterone concentrations, and a decreasing risk associated with increasing SHBG concentrations in women who were not randomly allocated to receive anastrozole,” they said.

“However, to our knowledge, this is the first report of the effect of low concentrations of estradiol or testosterone on a lack of response to aromatase inhibitor treatment, either as a preventive measure or in the adjuvant setting,” they added. “These data provide support for the hypothesis that preventive therapy with an aromatase inhibitor is likely to be most effective for women with higher estradiol-to-SHBG ratios and, conversely, of little or no benefit for those with low estradiol-to-SHBG ratios.”

Thus, measurement of estradiol and SHBG concentrations might be helpful in making decisions about using inhibitors both for treatment and prevention, they continued, underscoring the importance of using assays sensitive enough to measure low estradiol concentrations in the plasma in postmenopausal women.

“We used a very sensitive liquid chromatography–tandem mass spectroscopy assay (lower limit of sensitivity of 3 pmol/L), which allowed us to accurately measure the low concentrations of estradiol and SHBG in the serum samples from our population of postmenopausal women. Wider use of this type of assay or a similar assay will be necessary to implement any of the actions suggested by this study,” they explained.

The findings “suggest a potential role for measuring estradiol, testosterone, and SHBG more widely, both in determining which individuals are at high risk and the likely response to endocrine treatment,” they concluded, noting that measuring serum hormones is inexpensive and, if used more routinely in high-risk clinics and for treatment of early breast cancer, could “substantially improve disease management.”

This study was funded by Cancer Research UK, National Health and Medical Research Council (Australia), Breast Cancer Research Foundation, and DaCosta Fund. Dr. Cuzick reported receiving royalties from Cancer Research UK for commercial use of the IBIS (Tyrer-Cuzick) breast cancer risk evaluation software.

Measuring serum hormone concentrations in postmenopausal women at high risk for developing breast cancer may help identify those most likely to benefit from preventive treatment with an aromatase inhibitor, according to findings from a case-control study using data from a large breast cancer prevention trial.

In the randomized, placebo-controlled IBIS-II prevention trial of 3864 women aged 40-70 years at increased risk for developing breast cancer, treatment with the aromatase inhibitor anastrozole was associated with a 49% reduction in breast cancer incidence. At median follow-up of 131 months, breast cancer occurred in 85 (4.4%) versus 165 (8.5%) of patients in the anastrozole and placebo arms, respectively.

A preplanned case-control study involving 212 participants from the anastrozole group (72 cases and 140 controls) and 416 from the placebo group (142 cases and 274 controls), showed a significant trend toward increasing breast cancer risk with increasing estradiol-to-sex hormone binding globulin (SHBG) ratio in the placebo group, but not in the anastrozole group (trend per quartile, 1.25 vs 1.06), reported Jack Cuzick, PhD, of the Wolfson Institute of Population Health, Queen Mary University of London, UK, and colleagues.

A weaker but still significant effect was observed for the testosterone-SHBG ratio in the placebo group (trend, 1.21), but again, no such effect was seen in the anastrozole group (trend, 1.18).

A relative benefit was seen for anastrozole in estradiol concentration quartiles 2, 3, and 4 (relative risk [RR], 0.55, 0.54, and 0.56, respectively), but not in quartile 1.

The findings were published online December 6 in The Lancet Oncology.

Study participants were recruited from 153 breast cancer treatment centers across 18 countries and randomized in a 1:1 ratio to receive 1 mg of oral anastrozole daily or placebo. For the case-control analysis, the investigators looked at the effects of baseline estradiol to SHBG ratio on the development of all breast cancers, including ductal carcinoma in situ. They also assessed the relative benefit of anastrozole versus placebo.

Case patients were those diagnosed with breast cancer after trial entry through data cutoff on October 22, 2019, and who had not used hormone replacement therapy within 3 months of trial entry or during the trial. Controls were participants without breast cancer who were randomly selected and matched according to treatment group, age, and follow-up time.

“Although the association between estradiol and breast cancer risk is well established, less is known about whether the concentrations of these hormones have an effect on the efficacy of preventive therapy with selective estrogen receptor modulators or aromatase inhibitors in women at increased risk of developing breast cancer,” the investigators noted, explaining that in the current analysis, they “tested the hypothesis that, for women with a low estradiol–SHBG ratio, anastrozole would provide little or no reduction in the risk of breast cancer.”

The results from the placebo group “confirm the increasing risk of breast cancer associated with higher estradiol and testosterone concentrations, and a decreasing risk associated with increasing SHBG concentrations in women who were not randomly allocated to receive anastrozole,” they said.

“However, to our knowledge, this is the first report of the effect of low concentrations of estradiol or testosterone on a lack of response to aromatase inhibitor treatment, either as a preventive measure or in the adjuvant setting,” they added. “These data provide support for the hypothesis that preventive therapy with an aromatase inhibitor is likely to be most effective for women with higher estradiol-to-SHBG ratios and, conversely, of little or no benefit for those with low estradiol-to-SHBG ratios.”

Thus, measurement of estradiol and SHBG concentrations might be helpful in making decisions about using inhibitors both for treatment and prevention, they continued, underscoring the importance of using assays sensitive enough to measure low estradiol concentrations in the plasma in postmenopausal women.

“We used a very sensitive liquid chromatography–tandem mass spectroscopy assay (lower limit of sensitivity of 3 pmol/L), which allowed us to accurately measure the low concentrations of estradiol and SHBG in the serum samples from our population of postmenopausal women. Wider use of this type of assay or a similar assay will be necessary to implement any of the actions suggested by this study,” they explained.

The findings “suggest a potential role for measuring estradiol, testosterone, and SHBG more widely, both in determining which individuals are at high risk and the likely response to endocrine treatment,” they concluded, noting that measuring serum hormones is inexpensive and, if used more routinely in high-risk clinics and for treatment of early breast cancer, could “substantially improve disease management.”

This study was funded by Cancer Research UK, National Health and Medical Research Council (Australia), Breast Cancer Research Foundation, and DaCosta Fund. Dr. Cuzick reported receiving royalties from Cancer Research UK for commercial use of the IBIS (Tyrer-Cuzick) breast cancer risk evaluation software.

Measuring serum hormone concentrations in postmenopausal women at high risk for developing breast cancer may help identify those most likely to benefit from preventive treatment with an aromatase inhibitor, according to findings from a case-control study using data from a large breast cancer prevention trial.

In the randomized, placebo-controlled IBIS-II prevention trial of 3864 women aged 40-70 years at increased risk for developing breast cancer, treatment with the aromatase inhibitor anastrozole was associated with a 49% reduction in breast cancer incidence. At median follow-up of 131 months, breast cancer occurred in 85 (4.4%) versus 165 (8.5%) of patients in the anastrozole and placebo arms, respectively.

A preplanned case-control study involving 212 participants from the anastrozole group (72 cases and 140 controls) and 416 from the placebo group (142 cases and 274 controls), showed a significant trend toward increasing breast cancer risk with increasing estradiol-to-sex hormone binding globulin (SHBG) ratio in the placebo group, but not in the anastrozole group (trend per quartile, 1.25 vs 1.06), reported Jack Cuzick, PhD, of the Wolfson Institute of Population Health, Queen Mary University of London, UK, and colleagues.

A weaker but still significant effect was observed for the testosterone-SHBG ratio in the placebo group (trend, 1.21), but again, no such effect was seen in the anastrozole group (trend, 1.18).

A relative benefit was seen for anastrozole in estradiol concentration quartiles 2, 3, and 4 (relative risk [RR], 0.55, 0.54, and 0.56, respectively), but not in quartile 1.

The findings were published online December 6 in The Lancet Oncology.

Study participants were recruited from 153 breast cancer treatment centers across 18 countries and randomized in a 1:1 ratio to receive 1 mg of oral anastrozole daily or placebo. For the case-control analysis, the investigators looked at the effects of baseline estradiol to SHBG ratio on the development of all breast cancers, including ductal carcinoma in situ. They also assessed the relative benefit of anastrozole versus placebo.

Case patients were those diagnosed with breast cancer after trial entry through data cutoff on October 22, 2019, and who had not used hormone replacement therapy within 3 months of trial entry or during the trial. Controls were participants without breast cancer who were randomly selected and matched according to treatment group, age, and follow-up time.

“Although the association between estradiol and breast cancer risk is well established, less is known about whether the concentrations of these hormones have an effect on the efficacy of preventive therapy with selective estrogen receptor modulators or aromatase inhibitors in women at increased risk of developing breast cancer,” the investigators noted, explaining that in the current analysis, they “tested the hypothesis that, for women with a low estradiol–SHBG ratio, anastrozole would provide little or no reduction in the risk of breast cancer.”

The results from the placebo group “confirm the increasing risk of breast cancer associated with higher estradiol and testosterone concentrations, and a decreasing risk associated with increasing SHBG concentrations in women who were not randomly allocated to receive anastrozole,” they said.

“However, to our knowledge, this is the first report of the effect of low concentrations of estradiol or testosterone on a lack of response to aromatase inhibitor treatment, either as a preventive measure or in the adjuvant setting,” they added. “These data provide support for the hypothesis that preventive therapy with an aromatase inhibitor is likely to be most effective for women with higher estradiol-to-SHBG ratios and, conversely, of little or no benefit for those with low estradiol-to-SHBG ratios.”

Thus, measurement of estradiol and SHBG concentrations might be helpful in making decisions about using inhibitors both for treatment and prevention, they continued, underscoring the importance of using assays sensitive enough to measure low estradiol concentrations in the plasma in postmenopausal women.

“We used a very sensitive liquid chromatography–tandem mass spectroscopy assay (lower limit of sensitivity of 3 pmol/L), which allowed us to accurately measure the low concentrations of estradiol and SHBG in the serum samples from our population of postmenopausal women. Wider use of this type of assay or a similar assay will be necessary to implement any of the actions suggested by this study,” they explained.

The findings “suggest a potential role for measuring estradiol, testosterone, and SHBG more widely, both in determining which individuals are at high risk and the likely response to endocrine treatment,” they concluded, noting that measuring serum hormones is inexpensive and, if used more routinely in high-risk clinics and for treatment of early breast cancer, could “substantially improve disease management.”

This study was funded by Cancer Research UK, National Health and Medical Research Council (Australia), Breast Cancer Research Foundation, and DaCosta Fund. Dr. Cuzick reported receiving royalties from Cancer Research UK for commercial use of the IBIS (Tyrer-Cuzick) breast cancer risk evaluation software.

TOPLINE:

Removing race and incorporating social determinants of health (SDOH) into the pooled cohort risk equations (PCEs) for predicting atherosclerotic cardiovascular disease (ASCVD) outcomes made no difference to patients’ risk scores.

METHODOLOGY:

• Primary prevention guidelines recommend using risk prediction algorithms to assess the 10-year ASCVD risk, with the currently recommended PCEs including race.
• Researchers evaluated the incremental value of revised risk prediction equations excluding race and introducing SDOH in 11,638 participants from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) cohort.
• Participants were aged between 45 and 79 years, had no history of ASCVD, and were not taking statins.
• Participants were followed up to 10 years for incident ASCVD, including myocardial infarction, coronary heart disease death, and fatal and nonfatal stroke.

TAKEAWAY:

• Risk prediction equations performed similarly in race- and sex-stratified PCEs (C-statistic [95% CI])
• Black female: 0.71 (0.68-0.75); Black male: 0.68 (0.64-0.73); White female: 0.77 (0.74-0.80); White male: 0.68 (0.65-0.71)
• Race-free sex-specific PCEs yielded similar discrimination as race- and sex-stratified PCEs (C-statistic [95% CI]):
• Black female: 0.71 (0.67-0.75); Black male: 0.68 (0.63-0.72); White female: 0.76 (0.73-0.80); White male: 0.68 (0.65-0.71)
• The addition of SDOH to race-free sex-specific PCEs did not improve model performance (C-statistic [95% CI]):
• Black female: 0.72 (0.68-0.76); Black male: 0.68 (0.64-0.72); White female: 0.77 (0.74-0.80); White male: 0.68 (0.65-0.71)

IN PRACTICE:

“The major takeaway is we need to rethink the idea of race in cardiovascular risk prediction,” lead author Arnab Ghosh, MD, assistant professor of medicine at Weill Cornell Medical College and a hospitalist at New York-Presbyterian/Weill Cornell Medical Center, said in a press release.

“It’s essential for clinicians and scientists to consider how to appropriately address the health effects of race as a social construct, which has contributed to health disparities in cardiovascular outcomes,” Dr. Ghosh added.

SOURCE:

The study led by Dr. Ghosh was published online on December 6, 2023, in JAMA Cardiology with an Editor’s Note.

LIMITATIONS:

The study required informed consent for inclusion, which may have led to selection bias.

The REGARDS cohort’s SDOH may not have captured all social and socioeconomic influences on ASCVD outcomes.

DISCLOSURES:

The research was funded by the National Institute of Neurological Disorders and Stroke and the National Institute on Aging of the National Institutes of Health, Department of Health and Human Services, and others. Some authors declared receiving funding, grants, or personal fees from various sources.
 

A version of this article appeared on Medscape.com.

TOPLINE:

Removing race and incorporating social determinants of health (SDOH) into the pooled cohort risk equations (PCEs) for predicting atherosclerotic cardiovascular disease (ASCVD) outcomes made no difference to patients’ risk scores.

METHODOLOGY:

• Primary prevention guidelines recommend using risk prediction algorithms to assess the 10-year ASCVD risk, with the currently recommended PCEs including race.
• Researchers evaluated the incremental value of revised risk prediction equations excluding race and introducing SDOH in 11,638 participants from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) cohort.
• Participants were aged between 45 and 79 years, had no history of ASCVD, and were not taking statins.
• Participants were followed up to 10 years for incident ASCVD, including myocardial infarction, coronary heart disease death, and fatal and nonfatal stroke.

TAKEAWAY:

• Risk prediction equations performed similarly in race- and sex-stratified PCEs (C-statistic [95% CI])
• Black female: 0.71 (0.68-0.75); Black male: 0.68 (0.64-0.73); White female: 0.77 (0.74-0.80); White male: 0.68 (0.65-0.71)
• Race-free sex-specific PCEs yielded similar discrimination as race- and sex-stratified PCEs (C-statistic [95% CI]):
• Black female: 0.71 (0.67-0.75); Black male: 0.68 (0.63-0.72); White female: 0.76 (0.73-0.80); White male: 0.68 (0.65-0.71)
• The addition of SDOH to race-free sex-specific PCEs did not improve model performance (C-statistic [95% CI]):
• Black female: 0.72 (0.68-0.76); Black male: 0.68 (0.64-0.72); White female: 0.77 (0.74-0.80); White male: 0.68 (0.65-0.71)

IN PRACTICE:

“The major takeaway is we need to rethink the idea of race in cardiovascular risk prediction,” lead author Arnab Ghosh, MD, assistant professor of medicine at Weill Cornell Medical College and a hospitalist at New York-Presbyterian/Weill Cornell Medical Center, said in a press release.

“It’s essential for clinicians and scientists to consider how to appropriately address the health effects of race as a social construct, which has contributed to health disparities in cardiovascular outcomes,” Dr. Ghosh added.

SOURCE:

The study led by Dr. Ghosh was published online on December 6, 2023, in JAMA Cardiology with an Editor’s Note.

LIMITATIONS:

The study required informed consent for inclusion, which may have led to selection bias.

The REGARDS cohort’s SDOH may not have captured all social and socioeconomic influences on ASCVD outcomes.

DISCLOSURES:

The research was funded by the National Institute of Neurological Disorders and Stroke and the National Institute on Aging of the National Institutes of Health, Department of Health and Human Services, and others. Some authors declared receiving funding, grants, or personal fees from various sources.
 

A version of this article appeared on Medscape.com.

TOPLINE:

Removing race and incorporating social determinants of health (SDOH) into the pooled cohort risk equations (PCEs) for predicting atherosclerotic cardiovascular disease (ASCVD) outcomes made no difference to patients’ risk scores.

METHODOLOGY:

• Primary prevention guidelines recommend using risk prediction algorithms to assess the 10-year ASCVD risk, with the currently recommended PCEs including race.
• Researchers evaluated the incremental value of revised risk prediction equations excluding race and introducing SDOH in 11,638 participants from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) cohort.
• Participants were aged between 45 and 79 years, had no history of ASCVD, and were not taking statins.
• Participants were followed up to 10 years for incident ASCVD, including myocardial infarction, coronary heart disease death, and fatal and nonfatal stroke.

TAKEAWAY:

• Risk prediction equations performed similarly in race- and sex-stratified PCEs (C-statistic [95% CI])
• Black female: 0.71 (0.68-0.75); Black male: 0.68 (0.64-0.73); White female: 0.77 (0.74-0.80); White male: 0.68 (0.65-0.71)
• Race-free sex-specific PCEs yielded similar discrimination as race- and sex-stratified PCEs (C-statistic [95% CI]):
• Black female: 0.71 (0.67-0.75); Black male: 0.68 (0.63-0.72); White female: 0.76 (0.73-0.80); White male: 0.68 (0.65-0.71)
• The addition of SDOH to race-free sex-specific PCEs did not improve model performance (C-statistic [95% CI]):
• Black female: 0.72 (0.68-0.76); Black male: 0.68 (0.64-0.72); White female: 0.77 (0.74-0.80); White male: 0.68 (0.65-0.71)

IN PRACTICE:

“The major takeaway is we need to rethink the idea of race in cardiovascular risk prediction,” lead author Arnab Ghosh, MD, assistant professor of medicine at Weill Cornell Medical College and a hospitalist at New York-Presbyterian/Weill Cornell Medical Center, said in a press release.

“It’s essential for clinicians and scientists to consider how to appropriately address the health effects of race as a social construct, which has contributed to health disparities in cardiovascular outcomes,” Dr. Ghosh added.

SOURCE:

The study led by Dr. Ghosh was published online on December 6, 2023, in JAMA Cardiology with an Editor’s Note.

LIMITATIONS:

The study required informed consent for inclusion, which may have led to selection bias.

The REGARDS cohort’s SDOH may not have captured all social and socioeconomic influences on ASCVD outcomes.

DISCLOSURES:

The research was funded by the National Institute of Neurological Disorders and Stroke and the National Institute on Aging of the National Institutes of Health, Department of Health and Human Services, and others. Some authors declared receiving funding, grants, or personal fees from various sources.
 

A version of this article appeared on Medscape.com.

SAN DIEGO — Two proposed new multiple myeloma staging systems offer similar prognostic accuracy compared with the standard staging system, but the new systems provide more refined risk classifications across different disease stages. 

The findings should encourage greater use of these newer staging systems in routine clinical practice, first author Manni Mohyuddin, MD, said during a presentation at the American Society of Hematology annual meeting.

Dr. Mohyuddin and his colleagues retrospectively compared the standard Revised International Staging System (R-ISS) with two newer systems, the Second Revision of the R-ISS (R2-ISS) and the Mayo Additive Staging System (MASS), using real-world data from nearly 500 patients with newly diagnosed multiple myeloma.

The R-ISS, the most common multiple myeloma staging system, incorporates a range of prognostic features, including high-risk genetic markers assessed using fluorescence in situ hybridization as well as levels of lactate dehydrogenase, albumin, and beta-2 microglobulin, explained Dr. Mohyuddin, assistant professor at the Huntsman Cancer Institute, University of Utah, Salt Lake City.

R2-ISS and MASS include additional factors that reflect experts’ growing understanding of multiple myeloma. Specifically, the systems also evaluate a gain of chromosome 1q, in which patients have an extra copy of chromosome 1q, as well as the additive effects of multiple high-risk cytogenetic abnormalities, both of which indicate worse prognosis in multiple myeloma, Dr. Mohyuddin said in an interview.

To compare the three staging systems, the investigators used information on newly diagnosed patients in the Flatiron Health EHR–derived deidentified database, which includes data from cancer clinics across the United States. Patients were followed from first-line treatment initiation until death, the end of the study period, or last recorded activity.

The patients from the database had a median age of 70 years, and most had not received a transplant. The most common cytogenetic abnormality was gain 1q, present in about one third of patients. 

Given that the R2-ISS originated from patients in clinical trials, Dr. Mohyuddin noted the importance of assessing how the system would perform in a real-world setting. 

Of the 497 patients in the analysis, the R-ISS staging system classified 24% as stage I, 63% as stage II, and 13% as stage III. Overall survival differed across these R-ISS stages, indicating the system was prognostic for survival. Median overall survival was not reached for those with stage I disease, was 62.9 months for those with stage II disease, and 37.6 months for those with stage III disease.

Because the R-ISS doesn’t consider the additive effect of multiple cytogenetic abnormalities, many patients end up in the R-ISS stage II category but ultimately may have vastly different outcomes, Dr. Mohyuddin said.

The R2-ISS includes four risk categories, which provide more granularity to the stage II classification: Stage I is low risk, stage II is low-intermediate, stage III is intermediate, and stage IV is high risk. Using this staging system, 20% of patients were stage I, 25% were stage II, 46% were stage III, and 9% were stage IV.

The R2-ISS was also prognostic for survival, which generally worsened from stage I to stage IV: Median overall survival was not reached in stage I patients, was 69.3 months for stage II, 50.0 months for stage III, and 50.6 months for stage IV patients. However, Dr. Mohyuddin noted that there was some overlap in the survival curves for stages I and II and for stages III and IV.

When applying MASS, 34% of patients were categorized as stage I, 35% as stage II, and 31% as stage III disease. This system was prognostic for survival as well, with median overall survival of 76.9 months for stage I, 61.2 months for stage II, and 45.0 months for stage III.

With R2-ISS, many of those in R-ISS stage II are moved into stage I and III. With MASS, the R-ISS stage II patients are more evenly distributed across stages I, II, and III.

In other words, “we show that both these newer staging systems basically recategorize patients into different stages,” essentially “decreasing the number of people in the large, ambiguous (R-ISS) stage II category,” said Dr. Mohyuddin. 

Dr. Mohyuddin and colleagues also evaluated the staging systems in fully adjusted analyses that controlled for age, race/ethnicity, sex, practice type, and diagnosis year. 

Using R2-ISS, stage I patients had a similar risk for death compared with stage II patients (hazard ratio [HR], 1.2). Compared with stage I patients, stage III and IV patients had comparable risks for death, both about 2.5-fold higher than in those with stage I disease (HR, 2.4 and 2.6, respectively). 

Compared with stage I MASS patients, those with stage II had a twofold higher risk for death (HR, 2.0), and those with stage III had an almost threefold higher risk (HR, 2.7). 

Although no system considers all factors associated with myeloma outcomes, R2-ISS and MASS do offer a benefit over R-ISS, Dr. Mohyuddin said.

He added that the R2-ISS and MASS are similar from a statistical standpoint, but he gave MASS a slight edge for use in clinical practice.

MASS “more cleanly demarcated [patients] into prognostic subsets,” plus it is “a little easier to remember by heart,” he explained. MASS also puts more emphasis on the presence of multiple high-risk cytogenetic abnormalities, which is a worse prognostic in this era of quadruplet therapy for multiple myeloma, he added.

Because the study largely took place in an era when triplet therapy dominated, “we would be curious to see, with longer follow-up and more use of quadruplets, how these staging systems would perform,” he said. 

Despite the benefits of these newer staging systems, many factors play a role in multiple myeloma outcomes, Dr. Mohyuddin explained. Staging systems are “only a piece of the puzzle.”

Dr. Mohyuddin reported having no financial interests to disclose.
 

A version of this article appeared on Medscape.com.

SAN DIEGO — Two proposed new multiple myeloma staging systems offer similar prognostic accuracy compared with the standard staging system, but the new systems provide more refined risk classifications across different disease stages. 

The findings should encourage greater use of these newer staging systems in routine clinical practice, first author Manni Mohyuddin, MD, said during a presentation at the American Society of Hematology annual meeting.

Dr. Mohyuddin and his colleagues retrospectively compared the standard Revised International Staging System (R-ISS) with two newer systems, the Second Revision of the R-ISS (R2-ISS) and the Mayo Additive Staging System (MASS), using real-world data from nearly 500 patients with newly diagnosed multiple myeloma.

The R-ISS, the most common multiple myeloma staging system, incorporates a range of prognostic features, including high-risk genetic markers assessed using fluorescence in situ hybridization as well as levels of lactate dehydrogenase, albumin, and beta-2 microglobulin, explained Dr. Mohyuddin, assistant professor at the Huntsman Cancer Institute, University of Utah, Salt Lake City.

R2-ISS and MASS include additional factors that reflect experts’ growing understanding of multiple myeloma. Specifically, the systems also evaluate a gain of chromosome 1q, in which patients have an extra copy of chromosome 1q, as well as the additive effects of multiple high-risk cytogenetic abnormalities, both of which indicate worse prognosis in multiple myeloma, Dr. Mohyuddin said in an interview.

To compare the three staging systems, the investigators used information on newly diagnosed patients in the Flatiron Health EHR–derived deidentified database, which includes data from cancer clinics across the United States. Patients were followed from first-line treatment initiation until death, the end of the study period, or last recorded activity.

The patients from the database had a median age of 70 years, and most had not received a transplant. The most common cytogenetic abnormality was gain 1q, present in about one third of patients. 

Given that the R2-ISS originated from patients in clinical trials, Dr. Mohyuddin noted the importance of assessing how the system would perform in a real-world setting. 

Of the 497 patients in the analysis, the R-ISS staging system classified 24% as stage I, 63% as stage II, and 13% as stage III. Overall survival differed across these R-ISS stages, indicating the system was prognostic for survival. Median overall survival was not reached for those with stage I disease, was 62.9 months for those with stage II disease, and 37.6 months for those with stage III disease.

Because the R-ISS doesn’t consider the additive effect of multiple cytogenetic abnormalities, many patients end up in the R-ISS stage II category but ultimately may have vastly different outcomes, Dr. Mohyuddin said.

The R2-ISS includes four risk categories, which provide more granularity to the stage II classification: Stage I is low risk, stage II is low-intermediate, stage III is intermediate, and stage IV is high risk. Using this staging system, 20% of patients were stage I, 25% were stage II, 46% were stage III, and 9% were stage IV.

The R2-ISS was also prognostic for survival, which generally worsened from stage I to stage IV: Median overall survival was not reached in stage I patients, was 69.3 months for stage II, 50.0 months for stage III, and 50.6 months for stage IV patients. However, Dr. Mohyuddin noted that there was some overlap in the survival curves for stages I and II and for stages III and IV.

When applying MASS, 34% of patients were categorized as stage I, 35% as stage II, and 31% as stage III disease. This system was prognostic for survival as well, with median overall survival of 76.9 months for stage I, 61.2 months for stage II, and 45.0 months for stage III.

With R2-ISS, many of those in R-ISS stage II are moved into stage I and III. With MASS, the R-ISS stage II patients are more evenly distributed across stages I, II, and III.

In other words, “we show that both these newer staging systems basically recategorize patients into different stages,” essentially “decreasing the number of people in the large, ambiguous (R-ISS) stage II category,” said Dr. Mohyuddin. 

Dr. Mohyuddin and colleagues also evaluated the staging systems in fully adjusted analyses that controlled for age, race/ethnicity, sex, practice type, and diagnosis year. 

Using R2-ISS, stage I patients had a similar risk for death compared with stage II patients (hazard ratio [HR], 1.2). Compared with stage I patients, stage III and IV patients had comparable risks for death, both about 2.5-fold higher than in those with stage I disease (HR, 2.4 and 2.6, respectively). 

Compared with stage I MASS patients, those with stage II had a twofold higher risk for death (HR, 2.0), and those with stage III had an almost threefold higher risk (HR, 2.7). 

Although no system considers all factors associated with myeloma outcomes, R2-ISS and MASS do offer a benefit over R-ISS, Dr. Mohyuddin said.

He added that the R2-ISS and MASS are similar from a statistical standpoint, but he gave MASS a slight edge for use in clinical practice.

MASS “more cleanly demarcated [patients] into prognostic subsets,” plus it is “a little easier to remember by heart,” he explained. MASS also puts more emphasis on the presence of multiple high-risk cytogenetic abnormalities, which is a worse prognostic in this era of quadruplet therapy for multiple myeloma, he added.

Because the study largely took place in an era when triplet therapy dominated, “we would be curious to see, with longer follow-up and more use of quadruplets, how these staging systems would perform,” he said. 

Despite the benefits of these newer staging systems, many factors play a role in multiple myeloma outcomes, Dr. Mohyuddin explained. Staging systems are “only a piece of the puzzle.”

Dr. Mohyuddin reported having no financial interests to disclose.
 

A version of this article appeared on Medscape.com.

SAN DIEGO — Two proposed new multiple myeloma staging systems offer similar prognostic accuracy compared with the standard staging system, but the new systems provide more refined risk classifications across different disease stages. 

The findings should encourage greater use of these newer staging systems in routine clinical practice, first author Manni Mohyuddin, MD, said during a presentation at the American Society of Hematology annual meeting.

Dr. Mohyuddin and his colleagues retrospectively compared the standard Revised International Staging System (R-ISS) with two newer systems, the Second Revision of the R-ISS (R2-ISS) and the Mayo Additive Staging System (MASS), using real-world data from nearly 500 patients with newly diagnosed multiple myeloma.

The R-ISS, the most common multiple myeloma staging system, incorporates a range of prognostic features, including high-risk genetic markers assessed using fluorescence in situ hybridization as well as levels of lactate dehydrogenase, albumin, and beta-2 microglobulin, explained Dr. Mohyuddin, assistant professor at the Huntsman Cancer Institute, University of Utah, Salt Lake City.

R2-ISS and MASS include additional factors that reflect experts’ growing understanding of multiple myeloma. Specifically, the systems also evaluate a gain of chromosome 1q, in which patients have an extra copy of chromosome 1q, as well as the additive effects of multiple high-risk cytogenetic abnormalities, both of which indicate worse prognosis in multiple myeloma, Dr. Mohyuddin said in an interview.

To compare the three staging systems, the investigators used information on newly diagnosed patients in the Flatiron Health EHR–derived deidentified database, which includes data from cancer clinics across the United States. Patients were followed from first-line treatment initiation until death, the end of the study period, or last recorded activity.

The patients from the database had a median age of 70 years, and most had not received a transplant. The most common cytogenetic abnormality was gain 1q, present in about one third of patients. 

Given that the R2-ISS originated from patients in clinical trials, Dr. Mohyuddin noted the importance of assessing how the system would perform in a real-world setting. 

Of the 497 patients in the analysis, the R-ISS staging system classified 24% as stage I, 63% as stage II, and 13% as stage III. Overall survival differed across these R-ISS stages, indicating the system was prognostic for survival. Median overall survival was not reached for those with stage I disease, was 62.9 months for those with stage II disease, and 37.6 months for those with stage III disease.

Because the R-ISS doesn’t consider the additive effect of multiple cytogenetic abnormalities, many patients end up in the R-ISS stage II category but ultimately may have vastly different outcomes, Dr. Mohyuddin said.

The R2-ISS includes four risk categories, which provide more granularity to the stage II classification: Stage I is low risk, stage II is low-intermediate, stage III is intermediate, and stage IV is high risk. Using this staging system, 20% of patients were stage I, 25% were stage II, 46% were stage III, and 9% were stage IV.

The R2-ISS was also prognostic for survival, which generally worsened from stage I to stage IV: Median overall survival was not reached in stage I patients, was 69.3 months for stage II, 50.0 months for stage III, and 50.6 months for stage IV patients. However, Dr. Mohyuddin noted that there was some overlap in the survival curves for stages I and II and for stages III and IV.

When applying MASS, 34% of patients were categorized as stage I, 35% as stage II, and 31% as stage III disease. This system was prognostic for survival as well, with median overall survival of 76.9 months for stage I, 61.2 months for stage II, and 45.0 months for stage III.

With R2-ISS, many of those in R-ISS stage II are moved into stage I and III. With MASS, the R-ISS stage II patients are more evenly distributed across stages I, II, and III.

In other words, “we show that both these newer staging systems basically recategorize patients into different stages,” essentially “decreasing the number of people in the large, ambiguous (R-ISS) stage II category,” said Dr. Mohyuddin. 

Dr. Mohyuddin and colleagues also evaluated the staging systems in fully adjusted analyses that controlled for age, race/ethnicity, sex, practice type, and diagnosis year. 

Using R2-ISS, stage I patients had a similar risk for death compared with stage II patients (hazard ratio [HR], 1.2). Compared with stage I patients, stage III and IV patients had comparable risks for death, both about 2.5-fold higher than in those with stage I disease (HR, 2.4 and 2.6, respectively). 

Compared with stage I MASS patients, those with stage II had a twofold higher risk for death (HR, 2.0), and those with stage III had an almost threefold higher risk (HR, 2.7). 

Although no system considers all factors associated with myeloma outcomes, R2-ISS and MASS do offer a benefit over R-ISS, Dr. Mohyuddin said.

He added that the R2-ISS and MASS are similar from a statistical standpoint, but he gave MASS a slight edge for use in clinical practice.

MASS “more cleanly demarcated [patients] into prognostic subsets,” plus it is “a little easier to remember by heart,” he explained. MASS also puts more emphasis on the presence of multiple high-risk cytogenetic abnormalities, which is a worse prognostic in this era of quadruplet therapy for multiple myeloma, he added.

Because the study largely took place in an era when triplet therapy dominated, “we would be curious to see, with longer follow-up and more use of quadruplets, how these staging systems would perform,” he said. 

Despite the benefits of these newer staging systems, many factors play a role in multiple myeloma outcomes, Dr. Mohyuddin explained. Staging systems are “only a piece of the puzzle.”

Dr. Mohyuddin reported having no financial interests to disclose.
 

A version of this article appeared on Medscape.com.

TOPLINE:

In patients with Hashimoto disease and persistent symptoms despite adequate medical treatment, total thyroidectomy had a beneficial effect up to 5 years but with a substantially higher risk for complications than initially anticipated.

METHODOLOGY:

• The 5-year follow-up of 65 participants in a randomized, open-label trial of thyroidectomy plus medical management vs medical management alone aimed at testing the hypothesis that persistent symptoms despite adequate thyroxine replacement may be related to extrathyroidal autoimmune reactions and that complete removal of thyroid tissues may attenuate autoimmune responses and relieve symptoms.
• Patients in the control group were given the option of having surgery 18 months after enrollment, depending on trial results.
• The primary outcome was patient-reported health-related quality of life measured by the dimensional general health score in the generic Short Form-36 Health Survey questionnaire.

TAKEAWAY:

• The positive treatment effect seen after 18 months was maintained throughout the 3-year follow-up.
• In the intervention group, the improved general health score remained at the same level during the 5-year follow-up.
• Results were similar for the other Short Form-36 Health Survey domains and for total fatigue and chronic fatigue.
• Short-term (<12 months) or longer-lasting complications occurred in 23 patients, including 6 with recurrent laryngeal nerve paralysis (4 were long-term) and 12 with hypoparathyroidism (6 long-term, including 3 permanent).
• Five patients had postoperative hematoma and/or infection requiring intervention.

IN PRACTICE:

“The improvements in patient-reported outcome measures reported at 18 months after surgery were maintained at 5 years after surgery in the intervention group. In contrast, no spontaneous improvement was seen during 3 years in the control group.”

“Long-term complications in 10 of 73 (14%) patients despite use of meticulous dissection to achieve total thyroidectomy is unacceptably high. Medication and compensatory mechanisms for hypoparathyroidism and unilateral recurrent nerve injury, respectively, did alleviate symptoms.”

SOURCE:

This study was published in Annals of Internal Medicine, by Geir Hoff, MD, PhD, of the Department of Research, Telemark Hospital, Skien, and the Institute of Clinical Medicine, University of Oslo, Oslo, Norway, and colleagues.

LIMITATIONS:

None listed.

DISCLOSURES:

None.

TOPLINE:

In patients with Hashimoto disease and persistent symptoms despite adequate medical treatment, total thyroidectomy had a beneficial effect up to 5 years but with a substantially higher risk for complications than initially anticipated.

METHODOLOGY:

• The 5-year follow-up of 65 participants in a randomized, open-label trial of thyroidectomy plus medical management vs medical management alone aimed at testing the hypothesis that persistent symptoms despite adequate thyroxine replacement may be related to extrathyroidal autoimmune reactions and that complete removal of thyroid tissues may attenuate autoimmune responses and relieve symptoms.
• Patients in the control group were given the option of having surgery 18 months after enrollment, depending on trial results.
• The primary outcome was patient-reported health-related quality of life measured by the dimensional general health score in the generic Short Form-36 Health Survey questionnaire.

TAKEAWAY:

• The positive treatment effect seen after 18 months was maintained throughout the 3-year follow-up.
• In the intervention group, the improved general health score remained at the same level during the 5-year follow-up.
• Results were similar for the other Short Form-36 Health Survey domains and for total fatigue and chronic fatigue.
• Short-term (<12 months) or longer-lasting complications occurred in 23 patients, including 6 with recurrent laryngeal nerve paralysis (4 were long-term) and 12 with hypoparathyroidism (6 long-term, including 3 permanent).
• Five patients had postoperative hematoma and/or infection requiring intervention.

IN PRACTICE:

“The improvements in patient-reported outcome measures reported at 18 months after surgery were maintained at 5 years after surgery in the intervention group. In contrast, no spontaneous improvement was seen during 3 years in the control group.”

“Long-term complications in 10 of 73 (14%) patients despite use of meticulous dissection to achieve total thyroidectomy is unacceptably high. Medication and compensatory mechanisms for hypoparathyroidism and unilateral recurrent nerve injury, respectively, did alleviate symptoms.”

SOURCE:

This study was published in Annals of Internal Medicine, by Geir Hoff, MD, PhD, of the Department of Research, Telemark Hospital, Skien, and the Institute of Clinical Medicine, University of Oslo, Oslo, Norway, and colleagues.

LIMITATIONS:

None listed.

DISCLOSURES:

None.

TOPLINE:

In patients with Hashimoto disease and persistent symptoms despite adequate medical treatment, total thyroidectomy had a beneficial effect up to 5 years but with a substantially higher risk for complications than initially anticipated.

METHODOLOGY:

• The 5-year follow-up of 65 participants in a randomized, open-label trial of thyroidectomy plus medical management vs medical management alone aimed at testing the hypothesis that persistent symptoms despite adequate thyroxine replacement may be related to extrathyroidal autoimmune reactions and that complete removal of thyroid tissues may attenuate autoimmune responses and relieve symptoms.
• Patients in the control group were given the option of having surgery 18 months after enrollment, depending on trial results.
• The primary outcome was patient-reported health-related quality of life measured by the dimensional general health score in the generic Short Form-36 Health Survey questionnaire.

TAKEAWAY:

• The positive treatment effect seen after 18 months was maintained throughout the 3-year follow-up.
• In the intervention group, the improved general health score remained at the same level during the 5-year follow-up.
• Results were similar for the other Short Form-36 Health Survey domains and for total fatigue and chronic fatigue.
• Short-term (<12 months) or longer-lasting complications occurred in 23 patients, including 6 with recurrent laryngeal nerve paralysis (4 were long-term) and 12 with hypoparathyroidism (6 long-term, including 3 permanent).
• Five patients had postoperative hematoma and/or infection requiring intervention.

IN PRACTICE:

“The improvements in patient-reported outcome measures reported at 18 months after surgery were maintained at 5 years after surgery in the intervention group. In contrast, no spontaneous improvement was seen during 3 years in the control group.”

“Long-term complications in 10 of 73 (14%) patients despite use of meticulous dissection to achieve total thyroidectomy is unacceptably high. Medication and compensatory mechanisms for hypoparathyroidism and unilateral recurrent nerve injury, respectively, did alleviate symptoms.”

SOURCE:

This study was published in Annals of Internal Medicine, by Geir Hoff, MD, PhD, of the Department of Research, Telemark Hospital, Skien, and the Institute of Clinical Medicine, University of Oslo, Oslo, Norway, and colleagues.

LIMITATIONS:

None listed.

DISCLOSURES:

None.

Systematically biased artificial intelligence (AI) models did not improve clinicians’ accuracy in diagnosing hospitalized patients, based on data from more than 450 clinicians.

“Artificial Intelligence (AI) could support clinicians in their diagnostic decisions of hospitalized patients but could also be biased and cause potential harm,” said Sarah Jabbour, MSE, a PhD candidate in computer science and engineering at the University of Michigan, Ann Arbor, in an interview.

“Regulatory guidance has suggested that the use of AI explanations could mitigate these harms, but the effectiveness of using AI explanations has not been established,” she said.

To examine whether AI explanations can be effective in mitigating the potential harms of systemic bias in AI models, Ms. Jabbour and colleagues conducted a randomized clinical vignette survey study. The survey was administered between April 2022 and January 2023 across 13 states, and the study population included hospitalist physicians, nurse practitioners, and physician assistants. The results were published in JAMA.

Participants were randomized to AI predictions with AI explanations (226 clinicians) or without AI explanations (231 clinicians).

The primary outcome was diagnostic accuracy for pneumonia, heart failure, and chronic obstructive pulmonary disease, defined as the number of correct diagnoses over the total number of assessments, the researchers wrote.

The clinicians viewed nine clinical vignettes of patients hospitalized with acute respiratory failure, including their presenting symptoms, physical examination, laboratory results, and chest radiographs. Clinicians viewed two vignettes with no AI model input to establish baseline diagnostic accuracy. They made three assessments in each vignette, one for each diagnosis. The order of the vignettes was two without AI predictions (to establish baseline diagnostic accuracy), six with AI predictions, and one with a clinical consultation by a hypothetical colleague. The vignettes included standard and systematically biased AI models.

The baseline diagnostic accuracy was 73% for the diagnoses of pneumonia, heart failure, and chronic obstructive pulmonary disease. Clinicians’ accuracy increased by 2.9% when they viewed a standard diagnostic AI model without explanations and by 4.4% when they viewed models with AI explanations.

However, clinicians’ accuracy decreased by 11.3% after viewing systematically biased AI model predictions without explanations compared with baseline, and biased AI model predictions with explanations decreased accuracy by 9.1%.

The decrease in accuracy with systematically biased AI predictions without explanations was mainly attributable to a decrease in the participants’ diagnostic specificity, the researchers noted, but the addition of explanations did little to improve it, the researchers said.

Potentially Useful but Still Imperfect

The findings were limited by several factors including the use of a web-based survey, which differs from surveys in a clinical setting, the researchers wrote. Other limitations included the younger than average study population, and the focus on the clinicians making treatment decisions, vs other clinicians who might have a better understanding of the AI explanations.

“In our study, explanations were presented in a way that were considered to be obvious, where the AI model was completely focused on areas of the chest X-rays unrelated to the clinical condition,” Ms. Jabbour told this news organization. “We hypothesized that if presented with such explanations, the participants in our study would notice that the model was behaving incorrectly and not rely on its predictions. This was surprisingly not the case, and the explanations when presented alongside biased AI predictions had seemingly no effect in mitigating clinicians’ overreliance on biased AI,” she said.

“AI is being developed at an extraordinary rate, and our study shows that it has the potential to improve clinical decision-making. At the same time, it could harm clinical decision-making when biased,” Ms. Jabbour said. “We must be thoughtful about how to carefully integrate AI into clinical workflows, with the goal of improving clinical care while not introducing systematic errors or harming patients,” she added.

Looking ahead, “There are several potential research areas that could be explored,” said Ms. Jabbour. “Researchers should focus on careful validation of AI models to identify biased model behavior prior to deployment. AI researchers should also continue including and communicating with clinicians during the development of AI tools to better understand clinicians’ needs and how they interact with AI,” she said. “This is not an exhaustive list of research directions, and it will take much discussion between experts across disciplines such as AI, human computer interaction, and medicine to ultimately deploy AI safely into clinical care.”

Don’t Overestimate AI

“With the increasing use of artificial intelligence and machine learning in other spheres, there has been an increase in interest in exploring how they can be utilized to improve clinical outcomes,” said Suman Pal, MD, assistant professor in the division of hospital medicine at the University of New Mexico, Albuquerque, in an interview. “However, concerns remain regarding the possible harms and ways to mitigate them,” said Dr. Pal, who was not involved in the current study.

In the current study, “It was interesting to note that explanations did not significantly mitigate the decrease in clinician accuracy from systematically biased AI model predictions,” Dr. Pal said.

“For the clinician, the findings of this study caution against overreliance on AI in clinical decision-making, especially because of the risk of exacerbating existing health disparities due to systemic inequities in existing literature,” Dr. Pal told this news organization.

“Additional research is needed to explore how clinicians can be better trained in identifying both the utility and the limitations of AI and into methods of validation and continuous quality checks with integration of AI into clinical workflows,” he noted.

The study was funded by the National Heart, Lung, and Blood Institute. The researchers had no financial conflicts to disclose. Dr. Pal had no financial conflicts to disclose.

A version of this article first appeared on Medscape.com.

Systematically biased artificial intelligence (AI) models did not improve clinicians’ accuracy in diagnosing hospitalized patients, based on data from more than 450 clinicians.

“Artificial Intelligence (AI) could support clinicians in their diagnostic decisions of hospitalized patients but could also be biased and cause potential harm,” said Sarah Jabbour, MSE, a PhD candidate in computer science and engineering at the University of Michigan, Ann Arbor, in an interview.

“Regulatory guidance has suggested that the use of AI explanations could mitigate these harms, but the effectiveness of using AI explanations has not been established,” she said.

To examine whether AI explanations can be effective in mitigating the potential harms of systemic bias in AI models, Ms. Jabbour and colleagues conducted a randomized clinical vignette survey study. The survey was administered between April 2022 and January 2023 across 13 states, and the study population included hospitalist physicians, nurse practitioners, and physician assistants. The results were published in JAMA.

Participants were randomized to AI predictions with AI explanations (226 clinicians) or without AI explanations (231 clinicians).

The primary outcome was diagnostic accuracy for pneumonia, heart failure, and chronic obstructive pulmonary disease, defined as the number of correct diagnoses over the total number of assessments, the researchers wrote.

The clinicians viewed nine clinical vignettes of patients hospitalized with acute respiratory failure, including their presenting symptoms, physical examination, laboratory results, and chest radiographs. Clinicians viewed two vignettes with no AI model input to establish baseline diagnostic accuracy. They made three assessments in each vignette, one for each diagnosis. The order of the vignettes was two without AI predictions (to establish baseline diagnostic accuracy), six with AI predictions, and one with a clinical consultation by a hypothetical colleague. The vignettes included standard and systematically biased AI models.

The baseline diagnostic accuracy was 73% for the diagnoses of pneumonia, heart failure, and chronic obstructive pulmonary disease. Clinicians’ accuracy increased by 2.9% when they viewed a standard diagnostic AI model without explanations and by 4.4% when they viewed models with AI explanations.

However, clinicians’ accuracy decreased by 11.3% after viewing systematically biased AI model predictions without explanations compared with baseline, and biased AI model predictions with explanations decreased accuracy by 9.1%.

The decrease in accuracy with systematically biased AI predictions without explanations was mainly attributable to a decrease in the participants’ diagnostic specificity, the researchers noted, but the addition of explanations did little to improve it, the researchers said.

Potentially Useful but Still Imperfect

The findings were limited by several factors including the use of a web-based survey, which differs from surveys in a clinical setting, the researchers wrote. Other limitations included the younger than average study population, and the focus on the clinicians making treatment decisions, vs other clinicians who might have a better understanding of the AI explanations.

“In our study, explanations were presented in a way that were considered to be obvious, where the AI model was completely focused on areas of the chest X-rays unrelated to the clinical condition,” Ms. Jabbour told this news organization. “We hypothesized that if presented with such explanations, the participants in our study would notice that the model was behaving incorrectly and not rely on its predictions. This was surprisingly not the case, and the explanations when presented alongside biased AI predictions had seemingly no effect in mitigating clinicians’ overreliance on biased AI,” she said.

“AI is being developed at an extraordinary rate, and our study shows that it has the potential to improve clinical decision-making. At the same time, it could harm clinical decision-making when biased,” Ms. Jabbour said. “We must be thoughtful about how to carefully integrate AI into clinical workflows, with the goal of improving clinical care while not introducing systematic errors or harming patients,” she added.

Looking ahead, “There are several potential research areas that could be explored,” said Ms. Jabbour. “Researchers should focus on careful validation of AI models to identify biased model behavior prior to deployment. AI researchers should also continue including and communicating with clinicians during the development of AI tools to better understand clinicians’ needs and how they interact with AI,” she said. “This is not an exhaustive list of research directions, and it will take much discussion between experts across disciplines such as AI, human computer interaction, and medicine to ultimately deploy AI safely into clinical care.”

Don’t Overestimate AI

“With the increasing use of artificial intelligence and machine learning in other spheres, there has been an increase in interest in exploring how they can be utilized to improve clinical outcomes,” said Suman Pal, MD, assistant professor in the division of hospital medicine at the University of New Mexico, Albuquerque, in an interview. “However, concerns remain regarding the possible harms and ways to mitigate them,” said Dr. Pal, who was not involved in the current study.

In the current study, “It was interesting to note that explanations did not significantly mitigate the decrease in clinician accuracy from systematically biased AI model predictions,” Dr. Pal said.

“For the clinician, the findings of this study caution against overreliance on AI in clinical decision-making, especially because of the risk of exacerbating existing health disparities due to systemic inequities in existing literature,” Dr. Pal told this news organization.

“Additional research is needed to explore how clinicians can be better trained in identifying both the utility and the limitations of AI and into methods of validation and continuous quality checks with integration of AI into clinical workflows,” he noted.

The study was funded by the National Heart, Lung, and Blood Institute. The researchers had no financial conflicts to disclose. Dr. Pal had no financial conflicts to disclose.

A version of this article first appeared on Medscape.com.

Systematically biased artificial intelligence (AI) models did not improve clinicians’ accuracy in diagnosing hospitalized patients, based on data from more than 450 clinicians.

“Artificial Intelligence (AI) could support clinicians in their diagnostic decisions of hospitalized patients but could also be biased and cause potential harm,” said Sarah Jabbour, MSE, a PhD candidate in computer science and engineering at the University of Michigan, Ann Arbor, in an interview.

“Regulatory guidance has suggested that the use of AI explanations could mitigate these harms, but the effectiveness of using AI explanations has not been established,” she said.

To examine whether AI explanations can be effective in mitigating the potential harms of systemic bias in AI models, Ms. Jabbour and colleagues conducted a randomized clinical vignette survey study. The survey was administered between April 2022 and January 2023 across 13 states, and the study population included hospitalist physicians, nurse practitioners, and physician assistants. The results were published in JAMA.

Participants were randomized to AI predictions with AI explanations (226 clinicians) or without AI explanations (231 clinicians).

The primary outcome was diagnostic accuracy for pneumonia, heart failure, and chronic obstructive pulmonary disease, defined as the number of correct diagnoses over the total number of assessments, the researchers wrote.

The clinicians viewed nine clinical vignettes of patients hospitalized with acute respiratory failure, including their presenting symptoms, physical examination, laboratory results, and chest radiographs. Clinicians viewed two vignettes with no AI model input to establish baseline diagnostic accuracy. They made three assessments in each vignette, one for each diagnosis. The order of the vignettes was two without AI predictions (to establish baseline diagnostic accuracy), six with AI predictions, and one with a clinical consultation by a hypothetical colleague. The vignettes included standard and systematically biased AI models.

The baseline diagnostic accuracy was 73% for the diagnoses of pneumonia, heart failure, and chronic obstructive pulmonary disease. Clinicians’ accuracy increased by 2.9% when they viewed a standard diagnostic AI model without explanations and by 4.4% when they viewed models with AI explanations.

However, clinicians’ accuracy decreased by 11.3% after viewing systematically biased AI model predictions without explanations compared with baseline, and biased AI model predictions with explanations decreased accuracy by 9.1%.

The decrease in accuracy with systematically biased AI predictions without explanations was mainly attributable to a decrease in the participants’ diagnostic specificity, the researchers noted, but the addition of explanations did little to improve it, the researchers said.

Potentially Useful but Still Imperfect

The findings were limited by several factors including the use of a web-based survey, which differs from surveys in a clinical setting, the researchers wrote. Other limitations included the younger than average study population, and the focus on the clinicians making treatment decisions, vs other clinicians who might have a better understanding of the AI explanations.

“In our study, explanations were presented in a way that were considered to be obvious, where the AI model was completely focused on areas of the chest X-rays unrelated to the clinical condition,” Ms. Jabbour told this news organization. “We hypothesized that if presented with such explanations, the participants in our study would notice that the model was behaving incorrectly and not rely on its predictions. This was surprisingly not the case, and the explanations when presented alongside biased AI predictions had seemingly no effect in mitigating clinicians’ overreliance on biased AI,” she said.

“AI is being developed at an extraordinary rate, and our study shows that it has the potential to improve clinical decision-making. At the same time, it could harm clinical decision-making when biased,” Ms. Jabbour said. “We must be thoughtful about how to carefully integrate AI into clinical workflows, with the goal of improving clinical care while not introducing systematic errors or harming patients,” she added.

Looking ahead, “There are several potential research areas that could be explored,” said Ms. Jabbour. “Researchers should focus on careful validation of AI models to identify biased model behavior prior to deployment. AI researchers should also continue including and communicating with clinicians during the development of AI tools to better understand clinicians’ needs and how they interact with AI,” she said. “This is not an exhaustive list of research directions, and it will take much discussion between experts across disciplines such as AI, human computer interaction, and medicine to ultimately deploy AI safely into clinical care.”

Don’t Overestimate AI

“With the increasing use of artificial intelligence and machine learning in other spheres, there has been an increase in interest in exploring how they can be utilized to improve clinical outcomes,” said Suman Pal, MD, assistant professor in the division of hospital medicine at the University of New Mexico, Albuquerque, in an interview. “However, concerns remain regarding the possible harms and ways to mitigate them,” said Dr. Pal, who was not involved in the current study.

In the current study, “It was interesting to note that explanations did not significantly mitigate the decrease in clinician accuracy from systematically biased AI model predictions,” Dr. Pal said.

“For the clinician, the findings of this study caution against overreliance on AI in clinical decision-making, especially because of the risk of exacerbating existing health disparities due to systemic inequities in existing literature,” Dr. Pal told this news organization.

“Additional research is needed to explore how clinicians can be better trained in identifying both the utility and the limitations of AI and into methods of validation and continuous quality checks with integration of AI into clinical workflows,” he noted.

The study was funded by the National Heart, Lung, and Blood Institute. The researchers had no financial conflicts to disclose. Dr. Pal had no financial conflicts to disclose.

A version of this article first appeared on Medscape.com.

This transcript has been edited for clarity.

This is Dr. JoAnn Manson, professor of medicine at Harvard Medical School and Brigham and Women’s Hospital. I’d like to talk about a recent report in the journal Hypertension that raises questions about whether blood pressure (BP) guidelines should be revisited and whether sex-specific thresholds and targets should be considered. Current BP guidelines are sex-agnostic.

This study was done in the large-scale nationally representative NHANES cohort. It included more than 53,000 US men and women. The average age was about 45 years, with an average duration of follow-up of 9.5 years. During that time, about 2400 cardiovascular (CVD) deaths were documented at baseline. The BP was measured three times, and the results were averaged. About 20% of the cohort were taking antihypertensive medications, and 80% were not.

Sex differences were observed in the association between BP and CVD mortality. The systolic BP associated with the lowest risk for CVD death was 110-119 mm Hg in men and 100-109 mm Hg in women. In men, however, compared with a reference category of systolic BP of 100-109 mm Hg, the risk for CVD death began to increase significantly at a systolic BP ≥ 160 mm Hg, at which point, the hazard ratio was 1.76, or 76% higher risk.

In women, the risk for CVD death began to increase significantly at a lower threshold. Compared with a reference category of systolic BP of 100-109 mm Hg, women whose systolic BP was 130-139 mm Hg had a significant 61% increase in CVD death, and among those with a systolic BP of 140-159 mm Hg, the risk was increased by 75%. With a systolic BP ≥ 160 mm Hg, CVD deaths among women were more than doubled, with a hazard ratio of 2.13.

Overall, these findings suggest sex differences, with women having an increased risk for CVD death beginning at a lower elevation of their systolic BP. For diastolic BP, both men and women showed the typical U-shaped curve and the diastolic BP associated with the lowest risk for CVD death was 70-80 mm Hg.

If these findings can be replicated with additional research and other large-scale cohort studies, and randomized trials show differences in lowering BP, then sex-specific BP guidelines could have advantages and should be seriously considered. Furthermore, some of the CVD risk scores and risk modeling should perhaps use sex-specific blood pressure thresholds.Dr. Manson received study pill donation and infrastructure support from Mars Symbioscience (for the COSMOS trial).

A version of this article appeared on Medscape.com.

This transcript has been edited for clarity.

This is Dr. JoAnn Manson, professor of medicine at Harvard Medical School and Brigham and Women’s Hospital. I’d like to talk about a recent report in the journal Hypertension that raises questions about whether blood pressure (BP) guidelines should be revisited and whether sex-specific thresholds and targets should be considered. Current BP guidelines are sex-agnostic.

This study was done in the large-scale nationally representative NHANES cohort. It included more than 53,000 US men and women. The average age was about 45 years, with an average duration of follow-up of 9.5 years. During that time, about 2400 cardiovascular (CVD) deaths were documented at baseline. The BP was measured three times, and the results were averaged. About 20% of the cohort were taking antihypertensive medications, and 80% were not.

Sex differences were observed in the association between BP and CVD mortality. The systolic BP associated with the lowest risk for CVD death was 110-119 mm Hg in men and 100-109 mm Hg in women. In men, however, compared with a reference category of systolic BP of 100-109 mm Hg, the risk for CVD death began to increase significantly at a systolic BP ≥ 160 mm Hg, at which point, the hazard ratio was 1.76, or 76% higher risk.

In women, the risk for CVD death began to increase significantly at a lower threshold. Compared with a reference category of systolic BP of 100-109 mm Hg, women whose systolic BP was 130-139 mm Hg had a significant 61% increase in CVD death, and among those with a systolic BP of 140-159 mm Hg, the risk was increased by 75%. With a systolic BP ≥ 160 mm Hg, CVD deaths among women were more than doubled, with a hazard ratio of 2.13.

Overall, these findings suggest sex differences, with women having an increased risk for CVD death beginning at a lower elevation of their systolic BP. For diastolic BP, both men and women showed the typical U-shaped curve and the diastolic BP associated with the lowest risk for CVD death was 70-80 mm Hg.

If these findings can be replicated with additional research and other large-scale cohort studies, and randomized trials show differences in lowering BP, then sex-specific BP guidelines could have advantages and should be seriously considered. Furthermore, some of the CVD risk scores and risk modeling should perhaps use sex-specific blood pressure thresholds.Dr. Manson received study pill donation and infrastructure support from Mars Symbioscience (for the COSMOS trial).

A version of this article appeared on Medscape.com.

This transcript has been edited for clarity.

This is Dr. JoAnn Manson, professor of medicine at Harvard Medical School and Brigham and Women’s Hospital. I’d like to talk about a recent report in the journal Hypertension that raises questions about whether blood pressure (BP) guidelines should be revisited and whether sex-specific thresholds and targets should be considered. Current BP guidelines are sex-agnostic.

This study was done in the large-scale nationally representative NHANES cohort. It included more than 53,000 US men and women. The average age was about 45 years, with an average duration of follow-up of 9.5 years. During that time, about 2400 cardiovascular (CVD) deaths were documented at baseline. The BP was measured three times, and the results were averaged. About 20% of the cohort were taking antihypertensive medications, and 80% were not.

Sex differences were observed in the association between BP and CVD mortality. The systolic BP associated with the lowest risk for CVD death was 110-119 mm Hg in men and 100-109 mm Hg in women. In men, however, compared with a reference category of systolic BP of 100-109 mm Hg, the risk for CVD death began to increase significantly at a systolic BP ≥ 160 mm Hg, at which point, the hazard ratio was 1.76, or 76% higher risk.

In women, the risk for CVD death began to increase significantly at a lower threshold. Compared with a reference category of systolic BP of 100-109 mm Hg, women whose systolic BP was 130-139 mm Hg had a significant 61% increase in CVD death, and among those with a systolic BP of 140-159 mm Hg, the risk was increased by 75%. With a systolic BP ≥ 160 mm Hg, CVD deaths among women were more than doubled, with a hazard ratio of 2.13.

Overall, these findings suggest sex differences, with women having an increased risk for CVD death beginning at a lower elevation of their systolic BP. For diastolic BP, both men and women showed the typical U-shaped curve and the diastolic BP associated with the lowest risk for CVD death was 70-80 mm Hg.

If these findings can be replicated with additional research and other large-scale cohort studies, and randomized trials show differences in lowering BP, then sex-specific BP guidelines could have advantages and should be seriously considered. Furthermore, some of the CVD risk scores and risk modeling should perhaps use sex-specific blood pressure thresholds.Dr. Manson received study pill donation and infrastructure support from Mars Symbioscience (for the COSMOS trial).

A version of this article appeared on Medscape.com.

A study published last spring suggesting that hearing aids may help reduce dementia risk has been retracted due to a coding error identified by the authors. 

The study was published April 13 in The Lancet Public Health and reported at that time. It was retracted by the journal on December 12.

According to the retraction notice, the journal editors in late November were informed by the authors of the paper that an error was introduced in the output format setting of their SAS codes, which led to data for people with hearing loss using hearing aids and those with hearing loss without using hearing aids being switched. 

This led to errors in their analysis, “which render their findings and conclusions false and misleading,” the retraction notice states. 

These errors were identified by the researchers following an exchange with scientists seeking to reproduce the authors’ findings.In a statement, The Lancet Group said it “takes issues relating to research integrity extremely seriously” and follows best-practice guidance from the Committee on Publication Ethics (COPE) and the International Committee of Medical Journal Editors (ICMJE). 

“Retractions are a rare but important part of the publishing process, and we are grateful to the scientists who prompted the re-examination of the data,” the statement reads. 

Despite the retraction, other studies have suggested a link between hearing and dementia. 

One study of US Medicare beneficiaries found a 61% higher dementia prevalence in those with moderate to severe hearing loss compared to those with normal hearing.

In this research, even mild hearing loss was associated with increased dementia risk, although it was not statistically significant, and use of hearing aids was tied to a 32% decrease in dementia prevalence. 

In addition, a large meta-analysis showed that hearing aids significantly reduce the risk for cognitive decline and dementia and even improve short-term cognitive function in individuals with hearing loss.

A version of this article appeared on Medscape.com.

A study published last spring suggesting that hearing aids may help reduce dementia risk has been retracted due to a coding error identified by the authors. 

The study was published April 13 in The Lancet Public Health and reported at that time. It was retracted by the journal on December 12.

According to the retraction notice, the journal editors in late November were informed by the authors of the paper that an error was introduced in the output format setting of their SAS codes, which led to data for people with hearing loss using hearing aids and those with hearing loss without using hearing aids being switched. 

This led to errors in their analysis, “which render their findings and conclusions false and misleading,” the retraction notice states. 

These errors were identified by the researchers following an exchange with scientists seeking to reproduce the authors’ findings.In a statement, The Lancet Group said it “takes issues relating to research integrity extremely seriously” and follows best-practice guidance from the Committee on Publication Ethics (COPE) and the International Committee of Medical Journal Editors (ICMJE). 

“Retractions are a rare but important part of the publishing process, and we are grateful to the scientists who prompted the re-examination of the data,” the statement reads. 

Despite the retraction, other studies have suggested a link between hearing and dementia. 

One study of US Medicare beneficiaries found a 61% higher dementia prevalence in those with moderate to severe hearing loss compared to those with normal hearing.

In this research, even mild hearing loss was associated with increased dementia risk, although it was not statistically significant, and use of hearing aids was tied to a 32% decrease in dementia prevalence. 

In addition, a large meta-analysis showed that hearing aids significantly reduce the risk for cognitive decline and dementia and even improve short-term cognitive function in individuals with hearing loss.

A version of this article appeared on Medscape.com.

A study published last spring suggesting that hearing aids may help reduce dementia risk has been retracted due to a coding error identified by the authors. 

The study was published April 13 in The Lancet Public Health and reported at that time. It was retracted by the journal on December 12.

According to the retraction notice, the journal editors in late November were informed by the authors of the paper that an error was introduced in the output format setting of their SAS codes, which led to data for people with hearing loss using hearing aids and those with hearing loss without using hearing aids being switched. 

This led to errors in their analysis, “which render their findings and conclusions false and misleading,” the retraction notice states. 

These errors were identified by the researchers following an exchange with scientists seeking to reproduce the authors’ findings.In a statement, The Lancet Group said it “takes issues relating to research integrity extremely seriously” and follows best-practice guidance from the Committee on Publication Ethics (COPE) and the International Committee of Medical Journal Editors (ICMJE). 

“Retractions are a rare but important part of the publishing process, and we are grateful to the scientists who prompted the re-examination of the data,” the statement reads. 

Despite the retraction, other studies have suggested a link between hearing and dementia. 

One study of US Medicare beneficiaries found a 61% higher dementia prevalence in those with moderate to severe hearing loss compared to those with normal hearing.

In this research, even mild hearing loss was associated with increased dementia risk, although it was not statistically significant, and use of hearing aids was tied to a 32% decrease in dementia prevalence. 

In addition, a large meta-analysis showed that hearing aids significantly reduce the risk for cognitive decline and dementia and even improve short-term cognitive function in individuals with hearing loss.

A version of this article appeared on Medscape.com.

‘Twas the night before Festivus and all through the house, everyone was griping.

In case you’ve only been watching Friends reruns lately, Festivus is a holiday that originated 25 years ago in the last season of Seinfeld. George’s father created it as an alternative to Christmas hype. In addition to an aluminum pole, the holiday features the annual airing of grievances, when one is encouraged to voice complaints. Aluminum poles haven’t replaced Christmas trees, but the spirit of Festivus is still with us in the widespread airing of grievances in 2023.

Kaiser Permanente

Complaining isn’t just a post-pandemic problem. Hector spends quite a bit of time complaining about Paris in the Iliad. That was a few pandemics ago. And repining is ubiquitous in literature — as human as walking on two limbs it seems. Ostensibly, we complain to effect change: Something is wrong and we expect it to be different. But that’s not the whole story. No one believes the weather will improve or the Patriots will play better because we complain about them. So why do we bother?

Even if nothing changes on the outside, it does seem to alter our internal state, serving a healthy psychological function. Putting to words what is aggravating can have the same benefit of deep breathing. We describe it as “getting something off our chest” because that’s what it feels like. We feel unburdened just by saying it out loud. Complaining is also a way to bond with others. We have a strong instinct to be with people like ourselves and what better way to connect than to find common suffering? Think about the last time you complained: Cranky staff, prior auths, Medicare, disrespectful patients, many of your colleagues will nod in agreement, validating your feelings and making you feel less isolated.

There are also maladaptive reasons for whining. It’s obviously an elementary way to get attention or to remove responsibility. It can also be a political weapon (office politics included). It’s such a potent way to connect that it’s used to build alliances and clout. “Washington is doing a great job,” said no candidate ever. No, if you want to get people on your side, find something irritating and complain to everyone how annoying it is. This solidifies “us” versus “them,” which can harm organizations and families alike.

Yet, eliminating all complaints is neither feasible, nor probably advisable. You could try to make your office a complaint-free zone, but the likely result would be to push any griping to the remote corners where you can no longer hear them. These criticisms might have uncovered missed opportunities, identify problems, and even improve cohesion if done in a safe and transparent setting. If they are left unaddressed or if the underlying culture isn’t sound, then they can propagate and lead to factions that harm productivity.

Griping is as much part of the holiday season as jingle bells and jelly donuts. I don’t believe complaining is up now because people were grumpier in 2023. Rather I think people just craved connection more than ever. So join in: Traffic after the time change, Tesla service, (super) late patients, prior auths, perioral dermatitis, post-COVID telogen effluvium.

I feel better.

Dr. Benabio is director of Healthcare Transformation and chief of dermatology at Kaiser Permanente San Diego. The opinions expressed in this column are his own and do not represent those of Kaiser Permanente. Dr. Benabio is @Dermdoc on X (formerly Twitter). Write to him at [email protected].

‘Twas the night before Festivus and all through the house, everyone was griping.

In case you’ve only been watching Friends reruns lately, Festivus is a holiday that originated 25 years ago in the last season of Seinfeld. George’s father created it as an alternative to Christmas hype. In addition to an aluminum pole, the holiday features the annual airing of grievances, when one is encouraged to voice complaints. Aluminum poles haven’t replaced Christmas trees, but the spirit of Festivus is still with us in the widespread airing of grievances in 2023.

Kaiser Permanente

Complaining isn’t just a post-pandemic problem. Hector spends quite a bit of time complaining about Paris in the Iliad. That was a few pandemics ago. And repining is ubiquitous in literature — as human as walking on two limbs it seems. Ostensibly, we complain to effect change: Something is wrong and we expect it to be different. But that’s not the whole story. No one believes the weather will improve or the Patriots will play better because we complain about them. So why do we bother?

Even if nothing changes on the outside, it does seem to alter our internal state, serving a healthy psychological function. Putting to words what is aggravating can have the same benefit of deep breathing. We describe it as “getting something off our chest” because that’s what it feels like. We feel unburdened just by saying it out loud. Complaining is also a way to bond with others. We have a strong instinct to be with people like ourselves and what better way to connect than to find common suffering? Think about the last time you complained: Cranky staff, prior auths, Medicare, disrespectful patients, many of your colleagues will nod in agreement, validating your feelings and making you feel less isolated.

There are also maladaptive reasons for whining. It’s obviously an elementary way to get attention or to remove responsibility. It can also be a political weapon (office politics included). It’s such a potent way to connect that it’s used to build alliances and clout. “Washington is doing a great job,” said no candidate ever. No, if you want to get people on your side, find something irritating and complain to everyone how annoying it is. This solidifies “us” versus “them,” which can harm organizations and families alike.

Yet, eliminating all complaints is neither feasible, nor probably advisable. You could try to make your office a complaint-free zone, but the likely result would be to push any griping to the remote corners where you can no longer hear them. These criticisms might have uncovered missed opportunities, identify problems, and even improve cohesion if done in a safe and transparent setting. If they are left unaddressed or if the underlying culture isn’t sound, then they can propagate and lead to factions that harm productivity.

Griping is as much part of the holiday season as jingle bells and jelly donuts. I don’t believe complaining is up now because people were grumpier in 2023. Rather I think people just craved connection more than ever. So join in: Traffic after the time change, Tesla service, (super) late patients, prior auths, perioral dermatitis, post-COVID telogen effluvium.

I feel better.

Dr. Benabio is director of Healthcare Transformation and chief of dermatology at Kaiser Permanente San Diego. The opinions expressed in this column are his own and do not represent those of Kaiser Permanente. Dr. Benabio is @Dermdoc on X (formerly Twitter). Write to him at [email protected].

‘Twas the night before Festivus and all through the house, everyone was griping.

In case you’ve only been watching Friends reruns lately, Festivus is a holiday that originated 25 years ago in the last season of Seinfeld. George’s father created it as an alternative to Christmas hype. In addition to an aluminum pole, the holiday features the annual airing of grievances, when one is encouraged to voice complaints. Aluminum poles haven’t replaced Christmas trees, but the spirit of Festivus is still with us in the widespread airing of grievances in 2023.

Kaiser Permanente

Complaining isn’t just a post-pandemic problem. Hector spends quite a bit of time complaining about Paris in the Iliad. That was a few pandemics ago. And repining is ubiquitous in literature — as human as walking on two limbs it seems. Ostensibly, we complain to effect change: Something is wrong and we expect it to be different. But that’s not the whole story. No one believes the weather will improve or the Patriots will play better because we complain about them. So why do we bother?

Even if nothing changes on the outside, it does seem to alter our internal state, serving a healthy psychological function. Putting to words what is aggravating can have the same benefit of deep breathing. We describe it as “getting something off our chest” because that’s what it feels like. We feel unburdened just by saying it out loud. Complaining is also a way to bond with others. We have a strong instinct to be with people like ourselves and what better way to connect than to find common suffering? Think about the last time you complained: Cranky staff, prior auths, Medicare, disrespectful patients, many of your colleagues will nod in agreement, validating your feelings and making you feel less isolated.

There are also maladaptive reasons for whining. It’s obviously an elementary way to get attention or to remove responsibility. It can also be a political weapon (office politics included). It’s such a potent way to connect that it’s used to build alliances and clout. “Washington is doing a great job,” said no candidate ever. No, if you want to get people on your side, find something irritating and complain to everyone how annoying it is. This solidifies “us” versus “them,” which can harm organizations and families alike.

Yet, eliminating all complaints is neither feasible, nor probably advisable. You could try to make your office a complaint-free zone, but the likely result would be to push any griping to the remote corners where you can no longer hear them. These criticisms might have uncovered missed opportunities, identify problems, and even improve cohesion if done in a safe and transparent setting. If they are left unaddressed or if the underlying culture isn’t sound, then they can propagate and lead to factions that harm productivity.

Griping is as much part of the holiday season as jingle bells and jelly donuts. I don’t believe complaining is up now because people were grumpier in 2023. Rather I think people just craved connection more than ever. So join in: Traffic after the time change, Tesla service, (super) late patients, prior auths, perioral dermatitis, post-COVID telogen effluvium.

I feel better.

Dr. Benabio is director of Healthcare Transformation and chief of dermatology at Kaiser Permanente San Diego. The opinions expressed in this column are his own and do not represent those of Kaiser Permanente. Dr. Benabio is @Dermdoc on X (formerly Twitter). Write to him at [email protected].

Human microbiome research has progressed in leaps and bounds over the past decades, from pivotal studies begun in the 1970s to the launch of the Human Microbiome Project in 2007. Breakthroughs have laid the groundwork for more recent clinical applications, such as fecal microbiota transplantation (FMT), and advanced techniques to explore new therapeutic pathways. Yet the “microbiome revolution” is just getting started, according to professor Martin J. Blaser, MD, one of the field’s pioneers.

The ongoing research and clinical trials into the microbiome’s link to the major causes of death in the United States hold the promise of interventions that manipulate the microbiome to prevent, slow, or perhaps even cure these conditions, says Dr. Blaser, who holds the Henry Rutgers Chair of the Human Microbiome and is director of the Center for Advanced Biotechnology and Medicine at Rutgers University in New Brunswick, New Jersey.

Dr. Blaser is the author of Missing Microbes: How the Overuse of Antibiotics Is Fueling Our Modern Plagues, serves as chair of the Presidential Advisory Council on Combating Antibiotic-Resistant Bacteria and is a member of the scientific advisory board of the biotech startup Micronoma.

In this interview, which has been condensed and edited for clarity, Dr. Blaser discusses where we’re at now and where he sees the microbiome field evolving in the coming years.

Highlighting the Most Promising Applications

Which recent studies on the link between the human microbiome and disease have you found particularly promising?

There have been a number of studies, including our own, focusing on the gut-kidney axis. The gut microbiome produces, or detoxifies, metabolites that are toxic to the kidney: for example, those involved in the formation of kidney stones and in the worsening of uremia. 

Altering the microbiome to reduce the uremic toxins and the nidus for stone formation is a very promising field of research. 

What other disease states may be amenable to microbiome-based interventions?

There are diseases that are caused by known genetic mutations. Yet, for nearly all of them, there is great variation in clinical outcomes, which might be classed as genes multiplied by environment interactions. 

It seems likely to me that microbiome variation could account for some proportion of those differences for some genetic diseases. 

It’s now well established that altering the microbiome with FMT is a successful intervention for recurrent  Clostridioides difficile  infections. What do you see as the next disease states where FMT could prove successful?

If you go to ClinicalTrials.gov, you will find that that there are 471 trials registered using FMT. This is across a broad range of illnesses, including metabolic, immunological, autoimmune, inflammatory, degenerative, and neoplastic diseases. 

Which will be the next condition showing marked efficacy is anyone’s guess. That is why we must do clinical trials to assess what works and what does not, regardless of specific illness. 

The donor’s microbiome appears to be vital to engraftment success, with “superdonors” even being identified. What factors do you think primarily influence microbiome engraftment?

There is an emerging science about this question, driven in part by classical ecological theory. 

Right now, we are using FMT as if one size fits all. But this probably would not provide optimal treatment for all. Just as we type blood donors and recipients before the blood transfusion, one could easily imagine a parallel kind of procedure. 

Are there any diseases where it’s just too far-fetched to think altering the microbiome could make a difference?

The link between the microbiome and human health is so pervasive that there are few conditions that are out of the realm of possibility. It really is a frontier. 

Not that the microbiome causes everything, but by understanding and manipulating the microbiome, we could at least palliate, or slow down, particular pathologic processes. 

For all the major causes of death in the United States — cardiovascular disease, cancer, dementia and neurogenerative diseases, diabetes, and lung, liver, and kidney diseases — there is ongoing investigation of the microbiome. A greater promise would be to prevent or cure these illnesses. 

Predicting the Next Stages of the ‘Microbiome Revolution’

Do you believe we are at a turning point with the microbiome in terms of being able to manipulate or engineer it?

The microbiome is a scientific frontier that has an impact across the biosphere. It is a broad frontier involving human and veterinary medicine, agriculture, and the environment. Knowledge is increasing incrementally, as expected. 

Are we at the point yet where doctors should be incorporating microbiome-related lifestyle changes for people with or at risk for cancer, heart disease, Alzheimer’s disease, or other chronic conditions?

Although we are still in the early stages of the “microbiome revolution,” which I first wrote about in EMBO Reports  in 2006 and then again in the Journal of Clinical Investigation in 2014, I think important advances for all of these conditions are coming our way in the next 5-10 years. 

How are prebiotics, probiotics, and postbiotics being used to shape the microbiome?

This is a very important and active area in clinical investigation, which needs to be ramped up. 

Tens of millions of people are using probiotics and prebiotics every day for vague indications, and which have only infrequently been tested in robust clinical trials. So, there is a disconnect between what’s being claimed with the bulk of the probiotics at present and what we’ll actually know in the future. 

How do you think the microbiome will stack up to other factors influencing health, such as genetics, exercise, and nutrition?

All are important, but unlike genetics, the microbiome is tractable, like diet and exercise. 

It is essentially impossible to change one’s genome, but that might become more likely before too long. However, we can easily change someone’s microbiome through dietary means, for example. Once we know the ground rules, there will be many options. Right now, it is mostly one-offs, but as the scientific basis broadens, much more will be possible. 

In the future, do you think we’ll be able to look at a person’s microbiome and tell what his or her risk of developing disease is, similar to the way we use gene panels now?

Yes, but we will need scientific advances to teach us what are the important biomarkers in general and in particular people. This will be one area of precision medicine. 

Lessons From Decades at the Forefront

You’ve been involved in this research for over 30 years, and the majority has focused on the human microbiome and its role in disease. When did it become apparent to you that this research had unique therapeutic promise?

From the very start, there was always the potential to harness the microbiome to improve human health. In fact, I wrote a perspective in PNAS on that theme in 2010. 

The key is to understand the biology of the microbiome, and from the scientific study comes new preventives and new treatments. Right now, there are many “probiotic” products on the market. Probiotics have a great future, but most of what is out there has not been rigorously tested for effectiveness. 

Was there a particular series of studies that occurred before the launch of the Human Microbiome Project and brought us to the current era?

The studies in the 1970s-1980s by Carl Woese using 16S rRNA genes to understand phylogeny and evolution opened up the field of DNA sequencing to consider bacterial evolution and issues of ancestry. 

A key subject of your research and the focus of your book is antibiotic-resistant bacteria. What did this work teach you about describing the science of antibiotic resistance to the general public?

People don’t care very much about antibiotic resistance. They think that affects other people, mostly. In contrast, they care about their own health and their children’s health. 

The more that the data show that using antibiotics can be harmful to health in some circumstances, the more that use will diminish. We need more transparency about benefits and costs. 

Are there any common misconceptions about the microbiome that you hear from the general public, or even clinicians, that you would like to see greater efforts to dispel?

The public and the medical profession are in love with probiotics, buying them by the tens of millions. But as stated before, they are very diverse and mostly untested for efficacy. 

The next step is to test specific formulations to see which ones work, and for whom, and which ones don’t. That would be a big advance. 

A version of this article appeared on Medscape.com.

Human microbiome research has progressed in leaps and bounds over the past decades, from pivotal studies begun in the 1970s to the launch of the Human Microbiome Project in 2007. Breakthroughs have laid the groundwork for more recent clinical applications, such as fecal microbiota transplantation (FMT), and advanced techniques to explore new therapeutic pathways. Yet the “microbiome revolution” is just getting started, according to professor Martin J. Blaser, MD, one of the field’s pioneers.

The ongoing research and clinical trials into the microbiome’s link to the major causes of death in the United States hold the promise of interventions that manipulate the microbiome to prevent, slow, or perhaps even cure these conditions, says Dr. Blaser, who holds the Henry Rutgers Chair of the Human Microbiome and is director of the Center for Advanced Biotechnology and Medicine at Rutgers University in New Brunswick, New Jersey.

Dr. Blaser is the author of Missing Microbes: How the Overuse of Antibiotics Is Fueling Our Modern Plagues, serves as chair of the Presidential Advisory Council on Combating Antibiotic-Resistant Bacteria and is a member of the scientific advisory board of the biotech startup Micronoma.

In this interview, which has been condensed and edited for clarity, Dr. Blaser discusses where we’re at now and where he sees the microbiome field evolving in the coming years.

Highlighting the Most Promising Applications

Which recent studies on the link between the human microbiome and disease have you found particularly promising?

There have been a number of studies, including our own, focusing on the gut-kidney axis. The gut microbiome produces, or detoxifies, metabolites that are toxic to the kidney: for example, those involved in the formation of kidney stones and in the worsening of uremia. 

Altering the microbiome to reduce the uremic toxins and the nidus for stone formation is a very promising field of research. 

What other disease states may be amenable to microbiome-based interventions?

There are diseases that are caused by known genetic mutations. Yet, for nearly all of them, there is great variation in clinical outcomes, which might be classed as genes multiplied by environment interactions. 

It seems likely to me that microbiome variation could account for some proportion of those differences for some genetic diseases. 

It’s now well established that altering the microbiome with FMT is a successful intervention for recurrent  Clostridioides difficile  infections. What do you see as the next disease states where FMT could prove successful?

If you go to ClinicalTrials.gov, you will find that that there are 471 trials registered using FMT. This is across a broad range of illnesses, including metabolic, immunological, autoimmune, inflammatory, degenerative, and neoplastic diseases. 

Which will be the next condition showing marked efficacy is anyone’s guess. That is why we must do clinical trials to assess what works and what does not, regardless of specific illness. 

The donor’s microbiome appears to be vital to engraftment success, with “superdonors” even being identified. What factors do you think primarily influence microbiome engraftment?

There is an emerging science about this question, driven in part by classical ecological theory. 

Right now, we are using FMT as if one size fits all. But this probably would not provide optimal treatment for all. Just as we type blood donors and recipients before the blood transfusion, one could easily imagine a parallel kind of procedure. 

Are there any diseases where it’s just too far-fetched to think altering the microbiome could make a difference?

The link between the microbiome and human health is so pervasive that there are few conditions that are out of the realm of possibility. It really is a frontier. 

Not that the microbiome causes everything, but by understanding and manipulating the microbiome, we could at least palliate, or slow down, particular pathologic processes. 

For all the major causes of death in the United States — cardiovascular disease, cancer, dementia and neurogenerative diseases, diabetes, and lung, liver, and kidney diseases — there is ongoing investigation of the microbiome. A greater promise would be to prevent or cure these illnesses. 

Predicting the Next Stages of the ‘Microbiome Revolution’

Do you believe we are at a turning point with the microbiome in terms of being able to manipulate or engineer it?

The microbiome is a scientific frontier that has an impact across the biosphere. It is a broad frontier involving human and veterinary medicine, agriculture, and the environment. Knowledge is increasing incrementally, as expected. 

Are we at the point yet where doctors should be incorporating microbiome-related lifestyle changes for people with or at risk for cancer, heart disease, Alzheimer’s disease, or other chronic conditions?

Although we are still in the early stages of the “microbiome revolution,” which I first wrote about in EMBO Reports  in 2006 and then again in the Journal of Clinical Investigation in 2014, I think important advances for all of these conditions are coming our way in the next 5-10 years. 

How are prebiotics, probiotics, and postbiotics being used to shape the microbiome?

This is a very important and active area in clinical investigation, which needs to be ramped up. 

Tens of millions of people are using probiotics and prebiotics every day for vague indications, and which have only infrequently been tested in robust clinical trials. So, there is a disconnect between what’s being claimed with the bulk of the probiotics at present and what we’ll actually know in the future. 

How do you think the microbiome will stack up to other factors influencing health, such as genetics, exercise, and nutrition?

All are important, but unlike genetics, the microbiome is tractable, like diet and exercise. 

It is essentially impossible to change one’s genome, but that might become more likely before too long. However, we can easily change someone’s microbiome through dietary means, for example. Once we know the ground rules, there will be many options. Right now, it is mostly one-offs, but as the scientific basis broadens, much more will be possible. 

In the future, do you think we’ll be able to look at a person’s microbiome and tell what his or her risk of developing disease is, similar to the way we use gene panels now?

Yes, but we will need scientific advances to teach us what are the important biomarkers in general and in particular people. This will be one area of precision medicine. 

Lessons From Decades at the Forefront

You’ve been involved in this research for over 30 years, and the majority has focused on the human microbiome and its role in disease. When did it become apparent to you that this research had unique therapeutic promise?

From the very start, there was always the potential to harness the microbiome to improve human health. In fact, I wrote a perspective in PNAS on that theme in 2010. 

The key is to understand the biology of the microbiome, and from the scientific study comes new preventives and new treatments. Right now, there are many “probiotic” products on the market. Probiotics have a great future, but most of what is out there has not been rigorously tested for effectiveness. 

Was there a particular series of studies that occurred before the launch of the Human Microbiome Project and brought us to the current era?

The studies in the 1970s-1980s by Carl Woese using 16S rRNA genes to understand phylogeny and evolution opened up the field of DNA sequencing to consider bacterial evolution and issues of ancestry. 

A key subject of your research and the focus of your book is antibiotic-resistant bacteria. What did this work teach you about describing the science of antibiotic resistance to the general public?

People don’t care very much about antibiotic resistance. They think that affects other people, mostly. In contrast, they care about their own health and their children’s health. 

The more that the data show that using antibiotics can be harmful to health in some circumstances, the more that use will diminish. We need more transparency about benefits and costs. 

Are there any common misconceptions about the microbiome that you hear from the general public, or even clinicians, that you would like to see greater efforts to dispel?

The public and the medical profession are in love with probiotics, buying them by the tens of millions. But as stated before, they are very diverse and mostly untested for efficacy. 

The next step is to test specific formulations to see which ones work, and for whom, and which ones don’t. That would be a big advance. 

A version of this article appeared on Medscape.com.

Human microbiome research has progressed in leaps and bounds over the past decades, from pivotal studies begun in the 1970s to the launch of the Human Microbiome Project in 2007. Breakthroughs have laid the groundwork for more recent clinical applications, such as fecal microbiota transplantation (FMT), and advanced techniques to explore new therapeutic pathways. Yet the “microbiome revolution” is just getting started, according to professor Martin J. Blaser, MD, one of the field’s pioneers.

The ongoing research and clinical trials into the microbiome’s link to the major causes of death in the United States hold the promise of interventions that manipulate the microbiome to prevent, slow, or perhaps even cure these conditions, says Dr. Blaser, who holds the Henry Rutgers Chair of the Human Microbiome and is director of the Center for Advanced Biotechnology and Medicine at Rutgers University in New Brunswick, New Jersey.

Dr. Blaser is the author of Missing Microbes: How the Overuse of Antibiotics Is Fueling Our Modern Plagues, serves as chair of the Presidential Advisory Council on Combating Antibiotic-Resistant Bacteria and is a member of the scientific advisory board of the biotech startup Micronoma.

In this interview, which has been condensed and edited for clarity, Dr. Blaser discusses where we’re at now and where he sees the microbiome field evolving in the coming years.

Highlighting the Most Promising Applications

Which recent studies on the link between the human microbiome and disease have you found particularly promising?

There have been a number of studies, including our own, focusing on the gut-kidney axis. The gut microbiome produces, or detoxifies, metabolites that are toxic to the kidney: for example, those involved in the formation of kidney stones and in the worsening of uremia. 

Altering the microbiome to reduce the uremic toxins and the nidus for stone formation is a very promising field of research. 

What other disease states may be amenable to microbiome-based interventions?

There are diseases that are caused by known genetic mutations. Yet, for nearly all of them, there is great variation in clinical outcomes, which might be classed as genes multiplied by environment interactions. 

It seems likely to me that microbiome variation could account for some proportion of those differences for some genetic diseases. 

It’s now well established that altering the microbiome with FMT is a successful intervention for recurrent  Clostridioides difficile  infections. What do you see as the next disease states where FMT could prove successful?

If you go to ClinicalTrials.gov, you will find that that there are 471 trials registered using FMT. This is across a broad range of illnesses, including metabolic, immunological, autoimmune, inflammatory, degenerative, and neoplastic diseases. 

Which will be the next condition showing marked efficacy is anyone’s guess. That is why we must do clinical trials to assess what works and what does not, regardless of specific illness. 

The donor’s microbiome appears to be vital to engraftment success, with “superdonors” even being identified. What factors do you think primarily influence microbiome engraftment?

There is an emerging science about this question, driven in part by classical ecological theory. 

Right now, we are using FMT as if one size fits all. But this probably would not provide optimal treatment for all. Just as we type blood donors and recipients before the blood transfusion, one could easily imagine a parallel kind of procedure. 

Are there any diseases where it’s just too far-fetched to think altering the microbiome could make a difference?

The link between the microbiome and human health is so pervasive that there are few conditions that are out of the realm of possibility. It really is a frontier. 

Not that the microbiome causes everything, but by understanding and manipulating the microbiome, we could at least palliate, or slow down, particular pathologic processes. 

For all the major causes of death in the United States — cardiovascular disease, cancer, dementia and neurogenerative diseases, diabetes, and lung, liver, and kidney diseases — there is ongoing investigation of the microbiome. A greater promise would be to prevent or cure these illnesses. 

Predicting the Next Stages of the ‘Microbiome Revolution’

Do you believe we are at a turning point with the microbiome in terms of being able to manipulate or engineer it?

The microbiome is a scientific frontier that has an impact across the biosphere. It is a broad frontier involving human and veterinary medicine, agriculture, and the environment. Knowledge is increasing incrementally, as expected. 

Are we at the point yet where doctors should be incorporating microbiome-related lifestyle changes for people with or at risk for cancer, heart disease, Alzheimer’s disease, or other chronic conditions?

Although we are still in the early stages of the “microbiome revolution,” which I first wrote about in EMBO Reports  in 2006 and then again in the Journal of Clinical Investigation in 2014, I think important advances for all of these conditions are coming our way in the next 5-10 years. 

How are prebiotics, probiotics, and postbiotics being used to shape the microbiome?

This is a very important and active area in clinical investigation, which needs to be ramped up. 

Tens of millions of people are using probiotics and prebiotics every day for vague indications, and which have only infrequently been tested in robust clinical trials. So, there is a disconnect between what’s being claimed with the bulk of the probiotics at present and what we’ll actually know in the future. 

How do you think the microbiome will stack up to other factors influencing health, such as genetics, exercise, and nutrition?

All are important, but unlike genetics, the microbiome is tractable, like diet and exercise. 

It is essentially impossible to change one’s genome, but that might become more likely before too long. However, we can easily change someone’s microbiome through dietary means, for example. Once we know the ground rules, there will be many options. Right now, it is mostly one-offs, but as the scientific basis broadens, much more will be possible. 

In the future, do you think we’ll be able to look at a person’s microbiome and tell what his or her risk of developing disease is, similar to the way we use gene panels now?

Yes, but we will need scientific advances to teach us what are the important biomarkers in general and in particular people. This will be one area of precision medicine. 

Lessons From Decades at the Forefront

You’ve been involved in this research for over 30 years, and the majority has focused on the human microbiome and its role in disease. When did it become apparent to you that this research had unique therapeutic promise?

From the very start, there was always the potential to harness the microbiome to improve human health. In fact, I wrote a perspective in PNAS on that theme in 2010. 

The key is to understand the biology of the microbiome, and from the scientific study comes new preventives and new treatments. Right now, there are many “probiotic” products on the market. Probiotics have a great future, but most of what is out there has not been rigorously tested for effectiveness. 

Was there a particular series of studies that occurred before the launch of the Human Microbiome Project and brought us to the current era?

The studies in the 1970s-1980s by Carl Woese using 16S rRNA genes to understand phylogeny and evolution opened up the field of DNA sequencing to consider bacterial evolution and issues of ancestry. 

A key subject of your research and the focus of your book is antibiotic-resistant bacteria. What did this work teach you about describing the science of antibiotic resistance to the general public?

People don’t care very much about antibiotic resistance. They think that affects other people, mostly. In contrast, they care about their own health and their children’s health. 

The more that the data show that using antibiotics can be harmful to health in some circumstances, the more that use will diminish. We need more transparency about benefits and costs. 

Are there any common misconceptions about the microbiome that you hear from the general public, or even clinicians, that you would like to see greater efforts to dispel?

The public and the medical profession are in love with probiotics, buying them by the tens of millions. But as stated before, they are very diverse and mostly untested for efficacy. 

The next step is to test specific formulations to see which ones work, and for whom, and which ones don’t. That would be a big advance. 

A version of this article appeared on Medscape.com.

Patients with psoriasis being treated with biologics have a low overall risk for developing the paradoxical symptoms of eczema, with interleukin (IL)-23 inhibitors found to confer the lowest risk of the drug classes examined in a large observational analysis.

Using data from the British Association of Dermatologists Biologics and Immunomodulators Register (BADBIR) database, Ali Al-Janabi, MA, from the University of Manchester (England) and associates found that 273 (1%) of approximately 25,000 drug exposures in 13,699 biologic-treated patients with psoriasis were associated with paradoxical eczema.

The incidence of paradoxical eczema was found to vary by class. The highest rate was seen for IL-17 inhibitors, at 1.22 per 100,000 person-years, and the lowest rate was seen with IL-23 inhibitors, at 0.56 per 100,000 person-years. The respective incidence rates for tumor necrosis factor (TNF) inhibitors and IL-12/IL-23 inhibitors were a respective 0.94 and 0.80 per 100,000 person-years.

“Compared with TNF inhibitors, IL-23 inhibitor exposure was associated with significantly lower risk of paradoxical eczema,” the BADBIR Study Group reported in JAMA Dermatology. Indeed, patients treated with IL-23 inhibitors were 61% less likely than were those taking TNF-inhibitors to experience a paradoxical eczema event.

“These findings remained when restricting the analysis to first-line biologic exposures and were specific to this eczema phenotype” the group said.

Cautious Interpretation

As the corresponding author for the work, Mr. Al-Janabi observed in an email that the research needs to be replicated, and the findings need to be interpreted with caution.

“As well as usual clinical variables influencing biologic selection, clinicians could consider IL-23 inhibitors in patients with previous atopic dermatitis, hay fever, or paradoxical eczema episodes, as this class was associated with the lowest risk of paradoxical eczema,” he suggested.

A prior history of atopic dermatitis (AD) and hay fever appears to be particularly relevant, as both substantially upped the chances that paradoxical eczema would occur, with hazard ratios of 12.40 and 3.78, respectively. Increasing age also increased the risk, albeit slightly (hazard ratio [HR], 1.02 per year), and there was an apparent lower risk (HR, 0.60) comparing men and women.

The BADBIR Study Group authors believe that, to the best of their knowledge, this is the first study to compare paradoxical eczema risk by biologic class. “Based on clinical experience and prevalence of eczematous reactions reported in some IL-17 inhibitor clinical trials, we suspected an association between IL-17 inhibitor exposure and paradoxical eczema,” they wrote.

“While the incidence of paradoxical eczema was numerically highest among IL-17 inhibitor exposures, it was not significantly different from the incidence among TNF inhibitor exposures.” The low overall incidence of paradoxical eczema “may be reassuring for patients and clinicians,” they added, “but it is possible that the incidence was underestimated due to underreporting or exclusion of adverse events with insufficient detail.”

Details of the Analysis, Other Findings

To explore the risk of paradoxical eczema by biologic class and identify possible risk factors, the BADBIR Study Group performed a prospective cohort study using data held within the BADBIR database between September 2007 and December 2022.

Adults over the age of 18 year or older with plaque psoriasis and who had been treated with at least one of the following biologics were eligible for inclusion: the TNF inhibitors adalimumab, certolizumab pegol, etanercept, and infliximab; the IL-17 inhibitors bimekizumab, brodalumab, ixekizumab, and secukinumab; the IL-12/23 inhibitor ustekinumab; and the IL-23 inhibitors guselkumab, risankizumab, and tildrakizumab.

Patient records and adverse event data were reviewed to determine the incidence of paradoxical eczema events, using terms such as eczema, eczematized, eczematous, atopy, atopic, and dermatitis.

Of 24,952 drug exposures analyzed, the majority (11,819) were for TNF inhibitors, followed by IL-17 inhibitors (4,776), IL-12/23 inhibitors (6,423), and finally, IL-23 inhibitors (1,934).

Mr. Al-Janabi and coauthors reported that the median time to onset of paradoxical eczema events was 294 days — approximately 9.8 months. The earliest that these events were recorded was at 120 days (4 months), and the latest at 699 days (almost 2 years).

The face and neck were the most common sites affected (26% of exposures), with other sites including the limbs (23%), the trunk (13%), and hands or feet (12%). Itching (18%), redness (7%), and dryness (4%) were the most commonly reported symptoms.

The researchers noted that 21 patients had skin biopsies taken and “all showed spongiosis or a feature of eczema, with 1 having overlapping features of psoriasis.”

In the majority (92 %) of cases, patients experienced only one eczema event. Of the 20 patients who had more than one event, just over one-fifth of repeat events occurred after receiving the same biologic as for the index event. A quarter of events occurred after a different biologic of the same class had been used, and just over half of events occurred after a different class of biologic had been given.

Strengths and Limitations

The “large sample size and inclusion of multiple lines of exposure per participant” are strengths of the study, said the researchers. “We included data for all currently available biologics, originating from more than 160 dermatology centers in the UK and Ireland.”

They added, however, that the “main limitation is the small numbers of observations within certain subgroups, such as specific biologic exposures or participants in ethnic minority groups, restricting generalizability of our findings and the interpretation of some subgroup analyses.”

Moreover, the small number of paradoxical eczema events seen may have resulted in imprecise effect estimates, they observe, noting that the number of exposures to IL-23 inhibitors was low compared with other classes.

“Future studies with more exposures and paradoxical eczema events would enable a more robust analysis of individual drugs and patient subgroups,” the authors concluded.

The study was funded by the Medical Research Council. BADBIR is coordinated by The University of Manchester, and funded by the British Association of Dermatologists (BAD). The BAD receives income from AbbVie, Almirall, Amgen, Celgene, Janssen, LEO Pharma, Lilly, Novartis, Samsung Bioepis, Sandoz Hexal AG, and UCB Pharma for providing pharmacovigilance services. This income finances a separate contract between the BAD and The University of Manchester, which coordinates BADBIR. Mr. Al-Janabi reported receiving grants from the Medical Research Council during the conduct of the study; nonfinancial support from UCB, Almirall, and Janssen; and personal fees from UCB outside the submitted work.

Patients with psoriasis being treated with biologics have a low overall risk for developing the paradoxical symptoms of eczema, with interleukin (IL)-23 inhibitors found to confer the lowest risk of the drug classes examined in a large observational analysis.

Using data from the British Association of Dermatologists Biologics and Immunomodulators Register (BADBIR) database, Ali Al-Janabi, MA, from the University of Manchester (England) and associates found that 273 (1%) of approximately 25,000 drug exposures in 13,699 biologic-treated patients with psoriasis were associated with paradoxical eczema.

The incidence of paradoxical eczema was found to vary by class. The highest rate was seen for IL-17 inhibitors, at 1.22 per 100,000 person-years, and the lowest rate was seen with IL-23 inhibitors, at 0.56 per 100,000 person-years. The respective incidence rates for tumor necrosis factor (TNF) inhibitors and IL-12/IL-23 inhibitors were a respective 0.94 and 0.80 per 100,000 person-years.

“Compared with TNF inhibitors, IL-23 inhibitor exposure was associated with significantly lower risk of paradoxical eczema,” the BADBIR Study Group reported in JAMA Dermatology. Indeed, patients treated with IL-23 inhibitors were 61% less likely than were those taking TNF-inhibitors to experience a paradoxical eczema event.

“These findings remained when restricting the analysis to first-line biologic exposures and were specific to this eczema phenotype” the group said.

Cautious Interpretation

As the corresponding author for the work, Mr. Al-Janabi observed in an email that the research needs to be replicated, and the findings need to be interpreted with caution.

“As well as usual clinical variables influencing biologic selection, clinicians could consider IL-23 inhibitors in patients with previous atopic dermatitis, hay fever, or paradoxical eczema episodes, as this class was associated with the lowest risk of paradoxical eczema,” he suggested.

A prior history of atopic dermatitis (AD) and hay fever appears to be particularly relevant, as both substantially upped the chances that paradoxical eczema would occur, with hazard ratios of 12.40 and 3.78, respectively. Increasing age also increased the risk, albeit slightly (hazard ratio [HR], 1.02 per year), and there was an apparent lower risk (HR, 0.60) comparing men and women.

The BADBIR Study Group authors believe that, to the best of their knowledge, this is the first study to compare paradoxical eczema risk by biologic class. “Based on clinical experience and prevalence of eczematous reactions reported in some IL-17 inhibitor clinical trials, we suspected an association between IL-17 inhibitor exposure and paradoxical eczema,” they wrote.

“While the incidence of paradoxical eczema was numerically highest among IL-17 inhibitor exposures, it was not significantly different from the incidence among TNF inhibitor exposures.” The low overall incidence of paradoxical eczema “may be reassuring for patients and clinicians,” they added, “but it is possible that the incidence was underestimated due to underreporting or exclusion of adverse events with insufficient detail.”

Details of the Analysis, Other Findings

To explore the risk of paradoxical eczema by biologic class and identify possible risk factors, the BADBIR Study Group performed a prospective cohort study using data held within the BADBIR database between September 2007 and December 2022.

Adults over the age of 18 year or older with plaque psoriasis and who had been treated with at least one of the following biologics were eligible for inclusion: the TNF inhibitors adalimumab, certolizumab pegol, etanercept, and infliximab; the IL-17 inhibitors bimekizumab, brodalumab, ixekizumab, and secukinumab; the IL-12/23 inhibitor ustekinumab; and the IL-23 inhibitors guselkumab, risankizumab, and tildrakizumab.

Patient records and adverse event data were reviewed to determine the incidence of paradoxical eczema events, using terms such as eczema, eczematized, eczematous, atopy, atopic, and dermatitis.

Of 24,952 drug exposures analyzed, the majority (11,819) were for TNF inhibitors, followed by IL-17 inhibitors (4,776), IL-12/23 inhibitors (6,423), and finally, IL-23 inhibitors (1,934).

Mr. Al-Janabi and coauthors reported that the median time to onset of paradoxical eczema events was 294 days — approximately 9.8 months. The earliest that these events were recorded was at 120 days (4 months), and the latest at 699 days (almost 2 years).

The face and neck were the most common sites affected (26% of exposures), with other sites including the limbs (23%), the trunk (13%), and hands or feet (12%). Itching (18%), redness (7%), and dryness (4%) were the most commonly reported symptoms.

The researchers noted that 21 patients had skin biopsies taken and “all showed spongiosis or a feature of eczema, with 1 having overlapping features of psoriasis.”

In the majority (92 %) of cases, patients experienced only one eczema event. Of the 20 patients who had more than one event, just over one-fifth of repeat events occurred after receiving the same biologic as for the index event. A quarter of events occurred after a different biologic of the same class had been used, and just over half of events occurred after a different class of biologic had been given.

Strengths and Limitations

The “large sample size and inclusion of multiple lines of exposure per participant” are strengths of the study, said the researchers. “We included data for all currently available biologics, originating from more than 160 dermatology centers in the UK and Ireland.”

They added, however, that the “main limitation is the small numbers of observations within certain subgroups, such as specific biologic exposures or participants in ethnic minority groups, restricting generalizability of our findings and the interpretation of some subgroup analyses.”

Moreover, the small number of paradoxical eczema events seen may have resulted in imprecise effect estimates, they observe, noting that the number of exposures to IL-23 inhibitors was low compared with other classes.

“Future studies with more exposures and paradoxical eczema events would enable a more robust analysis of individual drugs and patient subgroups,” the authors concluded.

The study was funded by the Medical Research Council. BADBIR is coordinated by The University of Manchester, and funded by the British Association of Dermatologists (BAD). The BAD receives income from AbbVie, Almirall, Amgen, Celgene, Janssen, LEO Pharma, Lilly, Novartis, Samsung Bioepis, Sandoz Hexal AG, and UCB Pharma for providing pharmacovigilance services. This income finances a separate contract between the BAD and The University of Manchester, which coordinates BADBIR. Mr. Al-Janabi reported receiving grants from the Medical Research Council during the conduct of the study; nonfinancial support from UCB, Almirall, and Janssen; and personal fees from UCB outside the submitted work.

Patients with psoriasis being treated with biologics have a low overall risk for developing the paradoxical symptoms of eczema, with interleukin (IL)-23 inhibitors found to confer the lowest risk of the drug classes examined in a large observational analysis.

Using data from the British Association of Dermatologists Biologics and Immunomodulators Register (BADBIR) database, Ali Al-Janabi, MA, from the University of Manchester (England) and associates found that 273 (1%) of approximately 25,000 drug exposures in 13,699 biologic-treated patients with psoriasis were associated with paradoxical eczema.

The incidence of paradoxical eczema was found to vary by class. The highest rate was seen for IL-17 inhibitors, at 1.22 per 100,000 person-years, and the lowest rate was seen with IL-23 inhibitors, at 0.56 per 100,000 person-years. The respective incidence rates for tumor necrosis factor (TNF) inhibitors and IL-12/IL-23 inhibitors were a respective 0.94 and 0.80 per 100,000 person-years.

“Compared with TNF inhibitors, IL-23 inhibitor exposure was associated with significantly lower risk of paradoxical eczema,” the BADBIR Study Group reported in JAMA Dermatology. Indeed, patients treated with IL-23 inhibitors were 61% less likely than were those taking TNF-inhibitors to experience a paradoxical eczema event.

“These findings remained when restricting the analysis to first-line biologic exposures and were specific to this eczema phenotype” the group said.

Cautious Interpretation

As the corresponding author for the work, Mr. Al-Janabi observed in an email that the research needs to be replicated, and the findings need to be interpreted with caution.

“As well as usual clinical variables influencing biologic selection, clinicians could consider IL-23 inhibitors in patients with previous atopic dermatitis, hay fever, or paradoxical eczema episodes, as this class was associated with the lowest risk of paradoxical eczema,” he suggested.

A prior history of atopic dermatitis (AD) and hay fever appears to be particularly relevant, as both substantially upped the chances that paradoxical eczema would occur, with hazard ratios of 12.40 and 3.78, respectively. Increasing age also increased the risk, albeit slightly (hazard ratio [HR], 1.02 per year), and there was an apparent lower risk (HR, 0.60) comparing men and women.

The BADBIR Study Group authors believe that, to the best of their knowledge, this is the first study to compare paradoxical eczema risk by biologic class. “Based on clinical experience and prevalence of eczematous reactions reported in some IL-17 inhibitor clinical trials, we suspected an association between IL-17 inhibitor exposure and paradoxical eczema,” they wrote.

“While the incidence of paradoxical eczema was numerically highest among IL-17 inhibitor exposures, it was not significantly different from the incidence among TNF inhibitor exposures.” The low overall incidence of paradoxical eczema “may be reassuring for patients and clinicians,” they added, “but it is possible that the incidence was underestimated due to underreporting or exclusion of adverse events with insufficient detail.”

Details of the Analysis, Other Findings

To explore the risk of paradoxical eczema by biologic class and identify possible risk factors, the BADBIR Study Group performed a prospective cohort study using data held within the BADBIR database between September 2007 and December 2022.

Adults over the age of 18 year or older with plaque psoriasis and who had been treated with at least one of the following biologics were eligible for inclusion: the TNF inhibitors adalimumab, certolizumab pegol, etanercept, and infliximab; the IL-17 inhibitors bimekizumab, brodalumab, ixekizumab, and secukinumab; the IL-12/23 inhibitor ustekinumab; and the IL-23 inhibitors guselkumab, risankizumab, and tildrakizumab.

Patient records and adverse event data were reviewed to determine the incidence of paradoxical eczema events, using terms such as eczema, eczematized, eczematous, atopy, atopic, and dermatitis.

Of 24,952 drug exposures analyzed, the majority (11,819) were for TNF inhibitors, followed by IL-17 inhibitors (4,776), IL-12/23 inhibitors (6,423), and finally, IL-23 inhibitors (1,934).

Mr. Al-Janabi and coauthors reported that the median time to onset of paradoxical eczema events was 294 days — approximately 9.8 months. The earliest that these events were recorded was at 120 days (4 months), and the latest at 699 days (almost 2 years).

The face and neck were the most common sites affected (26% of exposures), with other sites including the limbs (23%), the trunk (13%), and hands or feet (12%). Itching (18%), redness (7%), and dryness (4%) were the most commonly reported symptoms.

The researchers noted that 21 patients had skin biopsies taken and “all showed spongiosis or a feature of eczema, with 1 having overlapping features of psoriasis.”

In the majority (92 %) of cases, patients experienced only one eczema event. Of the 20 patients who had more than one event, just over one-fifth of repeat events occurred after receiving the same biologic as for the index event. A quarter of events occurred after a different biologic of the same class had been used, and just over half of events occurred after a different class of biologic had been given.

Strengths and Limitations

The “large sample size and inclusion of multiple lines of exposure per participant” are strengths of the study, said the researchers. “We included data for all currently available biologics, originating from more than 160 dermatology centers in the UK and Ireland.”

They added, however, that the “main limitation is the small numbers of observations within certain subgroups, such as specific biologic exposures or participants in ethnic minority groups, restricting generalizability of our findings and the interpretation of some subgroup analyses.”

Moreover, the small number of paradoxical eczema events seen may have resulted in imprecise effect estimates, they observe, noting that the number of exposures to IL-23 inhibitors was low compared with other classes.

“Future studies with more exposures and paradoxical eczema events would enable a more robust analysis of individual drugs and patient subgroups,” the authors concluded.

The study was funded by the Medical Research Council. BADBIR is coordinated by The University of Manchester, and funded by the British Association of Dermatologists (BAD). The BAD receives income from AbbVie, Almirall, Amgen, Celgene, Janssen, LEO Pharma, Lilly, Novartis, Samsung Bioepis, Sandoz Hexal AG, and UCB Pharma for providing pharmacovigilance services. This income finances a separate contract between the BAD and The University of Manchester, which coordinates BADBIR. Mr. Al-Janabi reported receiving grants from the Medical Research Council during the conduct of the study; nonfinancial support from UCB, Almirall, and Janssen; and personal fees from UCB outside the submitted work.