© Todd Buchanan 2017
Poster session at ASH 2017 Photo courtesy of ASH
ATLANTA—A population-based study indicates that, compared to other cancer survivors, patients who survive diffuse large B-cell lymphoma (DLBCL) have an increased risk of autoimmune and infectious diseases.
For example, investigators found the risk of being diagnosed with impaired humoral immunity was 16.2 times higher in female DLBCL survivors than in breast cancer survivors, 14.8 times higher in male DLBCL survivors than in prostate cancer survivors, and 12.5 times higher in all DLBCL survivors than in survivors of head and neck cancer.
“Most of the treatments that we give for lymphoma have profound effects on the immune system, either directly or indirectly, including many of the T-cell-directed therapies,” said Tanaya Shree, MD, PhD, of Stanford University Medical Center in California.
“There have been studies on many of the effects suffered by lymphoma survivors, but very little is known about their immune health.”
Dr Shree and her colleagues undertook this study to determine how the immune system fares in lymphoma survivors. The investigators limited their analysis to survivors of DLBCL.
Dr Shree presented the findings at the 2017 ASH Annual Meeting (abstract 198*).
Study design
Investigators pulled data from the California Cancer Registry for patients with DLBCL as their first primary cancer diagnosed between 1991 and 2012. Patients had to be 18 or older at diagnosis and have survived more than a year after diagnosis.
“Importantly, we counted only diagnoses [of autoimmune and infectious diseases] that first appeared between 1 and 10 years after cancer diagnosis,” Dr Shree explained. “So any diagnosis we saw that had also been seen prior to cancer diagnosis or even up to 1 year post-cancer diagnosis, we considered to be pre-existing and were excluded from the analysis in order to really focus on new incident cases during survivorship.”
Investigators used the same criteria for the comparator cohorts.
The survivor data was linked to statewide discharge databases, and investigators performed the incidence analysis based on ICD-9 codes.
Investigators used Poisson regression analysis to obtain incident ratios and adjusted the models for age, race, and year of diagnosis.
They graphed the incident rate ratios for all the diagnoses that were significantly different between the DLBCL cohort and the comparator cohorts.
“[W]e considered a P value of less than 0.0005 to be significant,” Dr Shree clarified.
Survivor characteristics
The cohorts comprised 802,255 survivors of DLBCL (n=21,690), breast cancer (n=337,591), prostate cancer (n=325,533), melanoma (n=73,196), and head and neck cancer (n=44,245).
“At least 75% of patients in each cohort were aged 40 to 79,” Dr Shree noted, “with a good representation of elderly patients.”
The median follow-up time was 6.1 years for DLBCL patients and ranged from 5.7 years for head and neck cancer survivors to 8.3 years for prostate cancer survivors.
About three-quarters of patients in each cohort had hospitalization data within 1 to 10 years from cancer diagnosis.
DLBCL vs breast cancer
“Interestingly, we found some familiar names amongst the top-scoring diagnoses,” Dr Shree said.
Deficiency of humoral immunity (16.2-fold), autoimmune hemolytic anemia (9.9-fold), Sicca syndrome (6.9-fold), and immune thrombocytopenia (3.1-fold) were higher in female DLBCL survivors than breast cancer survivors.
“All of these have known associations with lymphoma,” Dr Shree said. “But we also found, surprisingly, increased rates of fungal [6.0-fold] and viral pneumonia [3.3-fold], and many other codes associated with respiratory infections. We also found a 3-fold increased rate of meningitis.”
“The only diagnosis statistically more common amongst breast cancer patients was cervicitis and endocervicitis, and this likely relates to the fact that many of these patients are undergoing hormone therapy.”
DLBCL vs prostate cancer
“We saw some of the same diagnoses come up as top-scoring hits, including viral [4.5-fold] and fungal pneumonia [8.2-fold], and meningitis [3.9-fold], and, in this case, Staphylococcal meningitis [8.6-fold],” Dr Shree said.
Deficiency of humoral immunity (14.8-fold), autoimmune hemolytic anemia (8.9-fold), Sicca syndrome (8.6-fold), and immune thrombocytopenia (4.8-fold) were also higher in the male DLBCL survivors than in prostate cancer survivors.
“No diagnoses were statistically more common in the prostate cancer survivors [than in male DLBCL survivors],” Dr Shree noted.
DLBCL vs head and neck cancer
“Again, the top 4 hits were the same 4 diagnoses we have been seeing repeatedly,” Dr Shree said.
Deficiency of humoral immunity (12.5-fold), autoimmune hemolytic anemia (9.3-fold), Sicca syndrome (5.5-fold), and immune thrombocytopenia (4.5-fold) were increased for DLBCL survivors compared to survivors of head and neck cancer.
DLBCL survivors also had an increased risk of respiratory infections, especially viral (4.4-fold) and fungal pneumonias (4.0-fold), meningitis (3.0-fold), and chronic lymphocytic thyroiditis (2.8-fold), also known as Hashimoto’s thyroiditis.
On the other hand, bacterial pneumonias and skin infections were more common in the head and neck cancer survivors than in DLBCL survivors.
DLBCL vs melanoma
“Interestingly, we did not see an increased risk for immune thrombocytopenias [in DLBCL survivors] compared to melanoma survivors in this comparison, which we had in all the other comparisons,” Dr Shree noted.
“But we did see [an increased risk for] the other diagnoses that we had been tracking, including, again, fungal pneumonia [6.9-fold], viral pneumonia [4.7-fold], and miscellaneous viral infections [2.6-fold].”
The only diagnosis that was statistically more common among melanoma survivors than DLBCL survivors was vitiligo.
Risks persist over time
The investigators assessed whether the elevated risks were static over the 1- to 10-year analysis period.
They took the top diagnoses—humoral deficiency, autoimmune hemolytic anemia, Sicca syndrome, and immune thrombocytopenia—and reviewed them for all cohorts to determine the rate of new cases.
“[F]or 3 out of these 4 diagnoses [humoral deficiency, autoimmune hemolytic anemia, and Sicca syndrome], increased incident rates are highest in the first 1 to 3 years after diagnosis in the lymphoma patients,” Dr Shree said.
“But even at 5 to 10 years out, these patients continue to have increased incidence of these diagnoses compared to the other cohorts, suggesting that these risks really do remain elevated over some time.”
The investigators repeated the analysis using broader categories of diagnoses with each category encompassing many ICD-9 codes.
“[I]n 12 out of 18 broad categories that we looked at, we can still find statistically significant differences in the incident rates for these diagnoses, and they were all increased in the lymphoma patients compared to the other cohorts,” Dr Shree explained.
“[T]hese increases were seen across multiple comparisons, suggesting that this phenomenon seems to be really lymphoma-specific and not specific to any of the individual comparisons we had chosen to perform.”
The findings, she said, have a lot of implications.
“We are particularly interested in which features of patients’ treatment contribute most to these elevated risks,” Dr Shree said. “And, of course, we want to know what to be able to tell our patients and how to follow them during survivorship.”
The investigators are currently validating their findings with further analysis of the Stanford lymphoma survivors cohort of approximately 3500 patients.
*Data in the abstract differ from the presentation.