Climate change influences on the annual onset of Lyme disease in the United States

Abstract

Lyme disease is the most commonly reported vector-borne illness in the United States. Lyme disease occurrence is highly seasonal and the annual springtime onset of cases is modulated by meteorological conditions in preceding months. A meteorological-based empirical model for Lyme disease onset week in the United States is driven with downscaled simulations from five global climate models and four greenhouse gas emissions scenarios to project the impacts of 21st century climate change on the annual onset week of Lyme disease. Projections are made individually and collectively for the 12 eastern States where >90% of cases occur. The national average annual onset week of Lyme disease is projected to become 0.4-0.5 weeks earlier for 2025-2040 (p<0.05), and 0.7-1.9 weeks earlier for 2065-2080 (p<0.01), with the largest shifts for scenarios with the highest greenhouse gas emissions. The more southerly mid-Atlantic States exhibit larger shifts (1.0-3.5 weeks) compared to the Northeastern and upper Midwestern States (0.2-2.3 weeks) by 2065-2080. Winter and spring temperature increases primarily cause the earlier onset. Greater spring precipitation and changes in humidity partially counteract the temperature effects. The model does not account for the possibility that abrupt shifts in the life cycle of Ixodes scapularis, the primary vector of the Lyme disease spirochete Borrelia burgdorferi in the eastern United States, may alter the disease transmission cycle in unforeseen ways. The results suggest 21st century climate change will make environmental conditions suitable for earlier annual onset of Lyme disease cases in the United States with possible implications for the timing of public health interventions.

Keywords: Borrelia burgdorferi; Climate change; Ixodes scapularis; Lyme disease.

Fig. 1

Box plots comparing the distributions of the national-level historical data for annual Lyme Onset Week (LOW) with the AOGCM multi-model mean distributions of LOW for each of the four RCP scenarios and two future periods. Each box plot shows the values of LOW for the maximum (top of dashed line), 75th percentile (top of box), mean (line through middle of box), 25th percentile (bottom of box) and minimum (bottom of dashed line) of the distribution. All distributions are comprised of values for 12 States and 16 years (N = 192). Circles along the top of each panel indicate whether the AOGCM multi-model mean is significantly different from the historical mean (see top legend). Box plot colors indicate different time periods (see middle legend). Black symbols on each box plot indicate the mean value of LOW from each individual AOGCM that contributes to the multi-model ensemble (see bottom legend).

Fig. 2

As in Fig. 1, but for the national-level climate variables including Jan-May average temperature (A), cumulative GDDs through Week 20 (B), average saturation deficit in the 5 weeks before LOW (C), and cumulative precipitation from week 8 until LOW (D).