Seasonal Differences in Incidence and Survival of Hodgkin Lymphoma Depend on Geographic Location of Diagnosis

It has been previously described that Hodgkin lymphoma (HL) exhibits a seasonal incidence and mortality pattern. However, these findings arise from small datasets and are partially conflicting. The aim of the present study is to provide the most comprehensive analysis of seasonal variation in incidence and mortality risk of HL with a focus on geographic differences in a very large, high-quality dataset.

All HL cases diagnosed between 1973 and 2012 in the 18 Surveillance, Epidemiology and End Results (SEER) registries were eligible (n=50,179). We excluded all death certificate and autopsy only cases (n=281), cases with unknown month of diagnosis (n=257), the "Alaska natives" (n=17) and "Rural Georgia" (n=57) registries due to low case numbers and cases with missing information on age at diagnosis (n=5) and Ann-arbor stage (n=8157), resulting in 41,405 cases being included in the study. All cases were grouped into quartiles according to the latitude of the case's county (Quartile 1: <34.19°N; Quartile 2: 34.19°N to 38.05°N; Quartile 3: 38.05°N to 41.68°N; Quartile 4: ≥41.68°N). The monthly incidence was adjusted for length of month. Cosinor analysis was employed to examine seasonality of incidence in all cases, latitude quartiles and other preplanned subgroups, such as gender, histological subtype, Ann-arbor stage and decade of diagnosis. The overall mortality risk within 3 years following HL diagnosis was analyzed using a Cox proportional-hazards model with a dichotomized variable representing season of diagnosis (Winter: Sept.-Feb., Summer: Mar.-Aug.). Known risk factors were corrected for.

HL exhibits a highly significant (p<0.001) seasonal incidence pattern with a peak in March and a trough in September. The adjusted incidence in September is approximately 15% higher than in March. This pattern was equally present in males (p<0.001) and females (p<0.001), but only present in the mixed cellularity (p<0.001), nodular sclerosis (p<0.001) and lymphocyte depleted subtypes (p=0.002). Seasonality was particularly pronounced in the age groups 20-29, 30-39 and 60-69, coinciding with age groups of increased HL incidence. All latitude quartiles showed significant seasonality patterns when analyzed separately, but the pattern was more pronounced in the two northern (≥38.05°N) quartiles with an amplitude of 0.102 ([0.073; 0.131], p<0.001) compared to the two southern (<38.05°N) quartiles with an amplitude of 0.055 ([0.026; 0.084], p<0.001). The difference was statistically significant (p=0.023).

Considering all cases, a diagnosis in winter does not significantly increase the risk of dying compared to being diagnosed in summer (HR=1.030 [0.981; 1.081], p=0.234). However, looking at cases diagnosed in the two northern (≥38.05°N) quartiles showed an increased risk of death within 3 years after HL diagnosis (HR=1.082 [1.009; 1.161], p=0.027), whereas no such relationship was evident for cases diagnosed in the two southern (<38.05°N) latitude quartiles (HR=0.990 [0.926; 1.059], p=0.772). To further examine the relationship between latitude and excessive mortality risk after a diagnosis made in winter, we included a multiplicative interaction term between the dichotomous season of diagnosis variable and latitude. The hazard ratio for the interaction term was 1.119 ([1.009; 1.241], p=0.033) indicating a significant interaction in the way that mortality risk after a HL diagnosis in winter compared to summer is only increased at higher latitudes. All results were confirmed using 5 years of follow-up data.

To conclude, HL exhibits a seasonal incidence and mortality pattern in this study, the most comprehensive analysis on this topic performed to date. The seasonal effect on the incidence of HL is exacerbated and the seasonal effect on mortality risk is only present at higher latitudes. Increased seasonal variations in Vitamin D levels at higher latitudes have been previously described. Therefore, a twofold protective effect of Vitamin D, where it both prevents HL development and increases the chance of survival after a diagnosis of HL is a possible explanation. Further studies, particularly those directly measuring Vitamin D levels in HL patients, are needed to improve the understanding of the role of Vitamin D in HL.