Hermida R C, Ayala D E
Bioengineering and Chronobiology Laboratories, E.T.S.I. Telecomunicación, University of Vigo, Spain.
Chronobiol Int. 1996 Oct;13(4):305-16. doi: 10.3109/07420529609020910.
A yearly pattern in the occurrence of uterine cervical cancer (UCC), obtained from cytological examinations reported as type V (cases concluding a malignant alteration), has been previously shown for data obtained in the Monterrey Metropolitan Area (state of Nuevo Leon, Mexico) for a span of 10 years (1978-1987), with a peak of relative incidence in the month of February being high stable for consecutive years. With the aim of extending and validating those results, we analyzed the monthly totals of positive detected cases of UCC in the states of Nuevo Leon, Chihuahua, Coahuila, and Tamaulipas (covering most of Northern Mexico) during the same period. To eliminate bias due to the seasonal variation in the number of screening smears, data were first expressed in relation to the number of cytological examinations done the same month. The least-squares fit of a 1-year cosine curve to the data of relative incidence in the four states reveals a statistically significant yearly pattern (p = .008), with a maximum of relative incidence in February almost double that during the rest of the year. Results indicate that the relative incidence of UCC is higher than the yearly average during the winter, with secondary peaks in May and October. In view of the nonsinusoidal waveform in the incidence of UCC, we undertook a multiple-component analysis, allowing several cosine functions to be simultaneously fitted to the data. Results indicate that the yearly pattern in the relative incidence of UCC can be represented by a model that includes two components with periods of 12 and 4 months (p = .004). The same model can be documented as statistically significant independently for each of the four states. These results, summarizing over 2200 positive cases of UCC detected in more than 1,100,000 screening smears, are in full agreement with those found previously for part of the state of Nuevo Leon and reveal a highly stable and predictable yearly pattern of variation in the relative incidence of UCC in Northern Mexico.
从报告为V型（诊断为恶性病变的病例）的细胞学检查中得出的子宫颈癌（UCC）发病的年度模式，先前已在墨西哥新莱昂州蒙特雷大都市区10年（1978 - 1987年）的数据中显示，相对发病率在2月份达到峰值，且连续数年保持高位稳定。为了扩展和验证这些结果，我们分析了同一时期新莱昂州、奇瓦瓦州、科阿韦拉州和塔毛利帕斯州（覆盖墨西哥北部大部分地区）UCC阳性检测病例的月度总数。为消除因筛查涂片数量季节性变化导致的偏差，数据首先按照同月进行的细胞学检查数量进行表述。对四个州相对发病率数据进行的1年余弦曲线最小二乘法拟合显示出具有统计学意义的年度模式（p = 0.008），2月份相对发病率最高，几乎是一年中其他时间的两倍。结果表明，UCC的相对发病率在冬季高于年度平均水平，在5月和10月出现次高峰。鉴于UCC发病率的波形并非正弦波，我们进行了多分量分析，允许几个余弦函数同时拟合数据。结果表明，UCC相对发病率的年度模式可以由一个包含周期为12个月和4个月的两个分量的模型表示（p = 0.004）。相同的模型在四个州中的每一个州都独立显示出具有统计学意义。这些结果总结了在超过110万次筛查涂片中检测到的2200多例UCC阳性病例，与先前在新莱昂州部分地区发现的结果完全一致，并揭示了墨西哥北部UCC相对发病率高度稳定且可预测的年度变化模式。
