Välimäki I A, Nieminen T, Antila K J, Southall D P
Cardiorespiratory Research Unit, University of Turku, Finland.
Ann N Y Acad Sci. 1988;533:228-37. doi: 10.1111/j.1749-6632.1988.tb37252.x.
In a prospective, population-based study, HRV was analyzed from 24-hr tape recordings made on 16 full-term and one preterm infant who had subsequently suffered SIDS and compared to similar data on 23 control infants (n of recordings, 44). In the SIDS group, heart rate was higher, and overall and beat-to-beat HRV (CV, CVS, respectively) were lower, than in the controls, but not significantly. Respiratory rate and respiratory HRV (by spectral analysis) were similar in both groups. Assuming that cardiorespiratory mechanisms of SIDS are multifactorial, we expected that several subgroups would be detected in both test groups. Therefore, the average data for each recording were subsequently examined by means of an expert system generator (ExTran, Intelligent Terminals Ltd., Edinburgh, UK). By rules induced with 25 nodes, the following results were obtained: 16/44 recordings were diagnosed as SIDS on the basis of (1) respiratory rate (RR) less than 33 and CV less than 3.46% (n = 8); (2) RR greater than 33, CVS less than 2.18%, and BW greater than 3,520 g (n = 4); and (3) RR greater than 33, CVS less than 2.18%, BW less than 3,520 g, HR greater than 136, and CV greater than 1.89% (n = 4). Seventeen of 44 were considered as non-SIDS when (1) RR was 33-47.4, CVS greater than 2.18%, and RSA less than 74.3 and (2) RR greater than 33, CVS less than 2.18%, BW less than 3,520 g, and HR less than 142. The remaining 11 cases required more complicated rules in order to be classified. This study shows that although the trend of increased HR and decreased HRV in the SIDS cases was statistically non-significant, an expert system program may be helpful in defining decision rules to identify cases of SIDS on the basis of cardiorespiratory data.
在一项基于人群的前瞻性研究中，对16名足月儿和1名早产儿（随后均死于婴儿猝死综合征）的24小时磁带记录进行心率变异性（HRV）分析，并与23名对照婴儿的类似数据（记录数量为44）进行比较。在婴儿猝死综合征组中，心率高于对照组，整体和逐搏HRV（分别为CV、CVS）低于对照组，但差异无统计学意义。两组的呼吸频率和呼吸HRV（通过频谱分析）相似。假设婴儿猝死综合征的心肺机制是多因素的，我们预期在两个测试组中都能检测到几个亚组。因此，随后通过专家系统生成器（ExTran，Intelligent Terminals Ltd.，爱丁堡，英国）检查每个记录的平均数据。通过由25个节点推导的规则，得到以下结果：44个记录中的16个被诊断为婴儿猝死综合征，依据为：（1）呼吸频率（RR）小于33且CV小于3.46%（n = 8）；（2）RR大于33，CVS小于2.18%，且出生体重（BW）大于3520克（n = 4）；以及（3）RR大于33，CVS小于2.18%，BW小于3520克，心率（HR）大于136，且CV大于1.89%（n = 4）。当（1）RR为33 - 47.4，CVS大于2.18%，且呼吸性窦性心律失常（RSA）小于74.3，以及（2）RR大于33，CVS小于2.18%，BW小于3520克，且HR小于142时，44个记录中的17个被视为非婴儿猝死综合征。其余11例需要更复杂的规则才能进行分类。这项研究表明，尽管婴儿猝死综合征病例中心率升高和HRV降低的趋势在统计学上无显著意义，但专家系统程序可能有助于定义决策规则，以便根据心肺数据识别婴儿猝死综合征病例。
