Evidence level: A
在气象条件几乎相同的两个年份里,地磁更活跃的那一年,呼吸道超额死亡是否更高?——这是把”天时”从”天气”里拆出来的关键一役。
正文
中医五运六气讲”天时”与”气候”共同决定疫病,但在现代统计里,”天时”(太阳活动与地磁扰动)和”气候”(温度、湿度)高度纠缠:地磁活跃的年份往往也是太阳活动的特定相位,而太阳活动又通过紫外线、气温间接影响流感流行。于是当研究者报告”地磁活跃期呼吸道死亡升高”时,批评者总可以反问一句:你拆得开吗?会不会只是气象条件的影子?
这个问题正是本系列此前一篇文章(#230,双通路拆分检验)留下的悬案:免疫通路(宿主昼夜节律、褪黑素)与气象通路(绝对湿度、温度)各有一批成熟证据,但从未在同一样本中把两条通路同时拆开。而”配对年份设计”(matched-year design)正是为这种质疑量身定做的减法实验。
配对年份设计的逻辑非常朴素:先不看地磁,只看气象。在二十年时间里,选出若干对”气象条件几乎相同”的年份——同样的冬季平均温度、同样的绝对湿度曲线、同样的流感流行强度。然后看每一对年份里,地磁更活跃的那一年,呼吸道超额死亡是否系统性更高。因为气象被”配对”掉了,剩下的差异只能归因于地磁。
2026年6月刚发表的一项全美研究,恰好把这个思路推进到了可执行的程度。Harfouch 等人用2000-2019年美国47个州的月度呼吸道感染(RTI)死亡数据,以地磁指数 Kp 的6个月移动平均作为暴露指标,用广义估计方程(GEE)泊松回归估计相对风险。结果发现:在1-12个月的所有暴露窗口中,6个月累积窗口与RTI死亡的关联最强;Kp 每升高一个四分位距,65-74岁、75-84岁、85岁以上三个年龄组的RTI死亡相对风险都显著升高(85岁以上女性 RR=1.16,95%CI 1.14-1.18;男性 RR=1.14,95%CI 1.12-1.16)。这是第一项在国家级尺度上把”地磁累积暴露”与”呼吸道感染死亡”直接挂钩的研究(PMID 42350784)。
为什么用6个月累积而不是当天Kp?这正是配对年份设计的灵魂所在。地磁对宿主的影响是慢性的、累积的——如果机制是昼夜节律紊乱和免疫调制,那么一次磁暴的当天效应很小,真正起作用的是连续几个月的”低强度持续扰动”。6个月窗口的峰值关联,与五运六气”客气加临、一气主六十日”的月尺度累积观,在时间结构上惊人地呼应:都是”一段时间的天时扰动”而非”一天的突发事件”。
但必须承认,PMID 42350784 是生态学研究,用的是全国月度汇总数据,温度、湿度只作为协变量进入模型,并没有做严格的”配对年份”。它的贡献是证明了”地磁→RTI死亡”关联在6个月窗口下稳健存在;而”边际增量”的直接检验——即气象完全匹配后地磁还能否贡献额外死亡——仍未有人做过。
这正是 #231 要立项的设计:气象条件配对年份 + 地磁活跃期事件研究法 + 6个月累积Kp暴露窗口 + 超额死亡估计。具体做法分四步:第一步,用绝对湿度(AH)和温度序列把2000-2025年划分成”气象相似年份组”(例如按冬季AH曲线的动态时间规整距离聚类);第二步,在每一组内比较地磁活跃年与平静年的RTI超额死亡(超额死亡用 Serfling 类周期回归基线估算,PMID 19453440、9209852 给出了四种成熟方法);第三步,用事件研究法(event study)检查地磁活跃期前后死亡率的时滞结构——如果免疫机制成立,峰值应出现在累积暴露之后数周到数月,而非当天;第四步,把6个月Kp累积窗口与其他窗口(1个月、12个月)做敏感性对比,确认窗口选择不是数据挖掘的产物。
这套设计的统计威力在于”配对”天然消除了气象混杂,而”超额死亡”天然吸收了年龄结构、慢性病负担等长期趋势。剩下的问题只有一个:地磁是否真的贡献了边际增量?
值得注意的是,太阳-流感关联本身是有争议的。Yeung 曾用1700-2000年的数据提出太阳黑子数可预测流感大流行(PMID 16806734),但 Towers 用更严格的统计方法重分析后认为,此前的阳性结论大多源于数据错误和任意分组,修正后没有显著关联(PMID 28847318)。这个对照非常重要:它说明”太阳-疾病”类关联极易被方法学瑕疵制造出来,配对年份设计恰恰是抵御这类虚假关联的防火墙——气象被固定后,地磁效应如果仍然出现,才配得上”边际增量”四个字。
证据与局限
已验证: 全美47州、2000-2019年数据显示,Kp指数6个月移动平均与RTI死亡显著相关,85岁以上人群RR最高(PMID 42350784);韩国2001-2019年数据显示日Kp与总死亡、心血管死亡相关,且调整PM10后仍稳健(PMID 40768856);263个美国城市2019年研究确认GMD与总死亡、心血管死亡相关(PMID 31511079)。
部分验证: 超额死亡估算方法成熟(Serfling周期回归、泊松回归、率差法四方法对比,PMID 19453440;121城市及时监测方法,PMID 9209852),但尚未有人将其与地磁暴露结合做”配对年份”检验;太阳-流感大流行关联存在正反两方证据(PMID 16806734 vs 28847318),结论依赖统计方法选择。
待验证: ①气象条件完全匹配后,地磁活跃期对RTI超额死亡的边际增量是否仍显著;②6个月累积Kp窗口是否优于其他窗口的机制解释(免疫累积 vs 气象代理);③配对年份设计能否在北美以外(如东亚、欧洲)复制,检验纬度与地磁强度梯度。
相关阅读
证据等级: A
参考文献: PMID 42350784, PMID 40768856, PMID 31511079, PMID 19453440, PMID 9209852, PMID 28847318, PMID 16806734
免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。
Marginal Increment of Geomagnetic Active Periods on Excess Respiratory Mortality: A Meteorological Condition Pairing Year Design
Evidence level: A
In two years with nearly identical meteorological conditions, is the year with higher geomagnetic activity associated with elevated excess respiratory mortality? This represents a critical step in disentangling “Heavenly Time” (solar activity and geomagnetic disturbance) from “Weather” (temperature, humidity).
Main Text
Traditional Chinese Medicine’s Five Movements and Six Qi posits that “Heavenly Time” and “climate” jointly determine epidemic outbreaks. However, in modern statistics, “Heavenly Time” (solar activity and geomagnetic disturbance) and “climate” (temperature, humidity) are highly entangled: years with high geomagnetic activity often coincide with specific phases of solar activity, which in turn indirectly influence influenza epidemics through ultraviolet radiation and temperature changes. Consequently, when researchers report elevated respiratory mortality during geomagnetic active periods, critics invariably ask: Can you disentangle them? Is it merely a shadow of meteorological conditions?
This question remains an open case from a previous article in this series (#230, Dual-Pathway Split Validation): while there is a robust body of evidence for both the immune pathway (host circadian rhythms, melatonin) and the meteorological pathway (absolute humidity, temperature), neither has simultaneously isolated both pathways within the same sample. The “matched-year design” is precisely the subtractive experiment tailored to address such skepticism.
The logic of the matched-year design is remarkably simple: first ignore geomagnetic activity and focus solely on meteorology. Over a twenty-year span, select pairs of years with “nearly identical meteorological conditions”—matching average winter temperatures, absolute humidity curves, and influenza epidemic intensity. Then, examine whether the year with higher geomagnetic activity within each pair shows systematically elevated excess respiratory mortality. Because meteorology is “matched out,” any remaining difference must be attributed to geomagnetism.
A study published in June 2026 across the United States advanced this approach to an executable level. Harfouch et al. utilized monthly respiratory tract infection (RTI) mortality data from 47 US states between 2000 and 2019, using a 6-month moving average of the geomagnetic index Kp as the exposure metric and generalized estimating equations (GEE) Poisson regression to estimate relative risk. The findings revealed that across all exposure windows from 1 to 12 months, the 6-month cumulative window showed the strongest association with RTI mortality; for every interquartile range increase in Kp, relative risks for RTI mortality significantly increased across three age groups (65-74, 75-84, and over 85 years; for females over 85, RR=1.16, 95%CI 1.14-1.18; for males, RR=1.14, 95%CI 1.12-1.16). This represents the first study at a national scale directly linking “geomagnetic cumulative exposure” to “respiratory infection mortality” (PMID 42350784).
Why use a 6-month cumulative window rather than daily Kp? This is the soul of the matched-year design. The effect of geomagnetism on hosts is chronic and cumulative—if the mechanism involves circadian rhythm disruption and immune modulation, the acute effect of a single geomagnetic storm is minimal; what truly functions is “low-intensity continuous disturbance” over consecutive months. The peak association at the 6-month window remarkably resonates with the monthly-scale cumulative view of “Guest Qi superposition and one Qi governing sixty days” in Five Movements and Six Qi: both represent “a period of Heavenly Time disturbance” rather than “a single-day突发事件”.
However, it must be acknowledged that PMID 42350784 is an ecological study using national monthly aggregated data, where temperature and humidity entered the model only as covariates without strict “matched-year” validation. Its contribution lies in demonstrating that the “geomagnetic → RTI mortality” association robustly exists under a 6-month window; however, the direct test of “marginal increment”—whether geomagnetism contributes additional mortality after meteorological conditions are fully matched—has yet to be conducted.
This is precisely the design proposed in #231: Meteorological Condition Matched-Year Design + Geomagnetic Active Period Event Study + 6-Month Cumulative Kp Exposure Window + Excess Mortality Estimation. The procedure involves four steps: First, partition 2000-2025 into “meteorologically similar year groups” using absolute humidity (AH) and temperature sequences (e.g., via dynamic time warping distance clustering); Second, compare RTI excess mortality between geomagnetic active years and quiet years within each group (excess mortality estimated using Serfling-type cyclical regression baselines, with PMID 19453440 and 9209852 outlining four mature methodologies); Third, employ an event study to examine the lag structure of mortality rates before and after geomagnetic active periods—if the immune mechanism holds, peaks should appear weeks or months after cumulative exposure rather than on the day; Fourth, conduct sensitivity comparisons between the 6-month Kp cumulative window and other windows (1 month, 12 months) to confirm the window selection is not a product of data dredging.
The statistical power of this design lies in how “matching” naturally eliminates meteorological confounding, while “excess mortality” inherently absorbs long-term trends such as age structure and chronic disease burden. The remaining question is singular: does geomagnetism truly contribute a marginal increment?
It is worth noting that the solar-influenza association itself remains controversial. Yeung proposed that sunspot numbers could predict influenza pandemics using data from 1700-2000 (PMID 16806734), yet Towers re-analyzed this using stricter statistical methods, concluding that prior positive findings largely stemmed from data errors and arbitrary grouping, yielding no significant association upon correction (PMID 28847318). This contrast is crucial: it demonstrates that “solar-disease” associations are easily manufactured by methodological flaws. The matched-year design serves as a firewall against such spurious associations; only if the geomagnetic effect persists after meteorological conditions are fixed does it warrant the term “marginal increment.”
Evidence and Limitations
Verified: National data from 47 US states (2000-2019) shows a significant correlation between the 6-month moving average of the Kp index and RTI mortality, with the highest relative risk observed in populations over 85 years old (PMID 42350784); Korean data from 2001-2019 indicates daily Kp correlates with total and cardiovascular mortality, remaining robust after adjusting for PM10 (PMID 40768856); a study of 263 US cities in 2019 confirmed correlations between GMD and total and cardiovascular mortality (PMID 31511079).
Partially Verified: Mature methodologies exist for excess mortality estimation (Serfling cyclical regression, Poisson regression, rate difference methods compared across four approaches, PMID 19453440; timely monitoring methods in 121 cities, PMID 9209852), yet no one has combined these with geomagnetic exposure for a “matched-year” validation; solar-influenza pandemic associations exhibit conflicting evidence (PMID 16806734 vs 28847318), with conclusions dependent on statistical methodology selection.
To Be Verified: ① Whether a marginal increment of geomagnetic active periods on RTI excess mortality remains significant after meteorological conditions are fully matched; ② Whether the 6-month cumulative Kp window offers superior mechanistic explanation compared to other windows (immune accumulation vs. meteorological proxy); ③ Whether the matched-year design can be replicated outside North America (e.g., East Asia, Europe) to test latitude and geomagnetic intensity gradients.
Related Reading
- Disentangling the Contributions of “Host Immune Pathway vs. Meteorological Condition Pathway” to Respiratory Virus Transmission via Geomagnetic Disturbance
- Sunspot Cycle and Epidemic Rhythm
- Research Pool: Spatio-Temporal Frequency Topology
Evidence Level: A
References: PMID 42350784, PMID 40768856, PMID 31511079, PMID 19453440, PMID 9209852, PMID 28847318, PMID 16806734
Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.
Leave a Reply