首先，你需要使用 s 使 . 符号搜索多行匹配。其次，不要使用 .* 因为在你的情况下，它总是只匹配第一次约会。第三，使用 preg_quote 从要查找的字符串中引用特殊字符。
作为一个完整的示例，请尝试以下方法：

// To avoid errors or unexpected results, we must escape regex characters
// For example, without this we cannot find some punctuation marks
$search = preg_quote('polis', '#');

// A simple regular expression to find the date
// Since we have to use it twice, we do not repeat ourselves
$date_rgxp = '\d{1,2}/\d{1,2}/\d{4}';

// Now it’s time to capture our date, for this we are using a more complex regex
// - Use "#" as delimiter, as it will allow do not escape "/" from the $date_rgxp
// - Use a named group to capture the date (it can be written as "($date_rgxp)", but in this case we should print it as "$m[1]" which is not so good for readability)
// - There can be any characters between the date and search word, but it should not contain another date (to solve this we are using negative lookahead "(?:(?!{$date_rgxp}).)+" instead of ".*")
// - Use the "si" modifiers to perform a case-insensitive multi-line search
if (preg_match("#(?P<date>{$date_rgxp})(?:(?!{$date_rgxp}).)+{$search}#si", $subject, $m)) {
    echo $m['date'];
}

下面是我用三个案例测试的工作代码：
案例1，前面的日期，你已经做了。
案例2，最后的日期，我只是对第一个做了一点修改。
案例3，日期在新行中，解释如下。 preg_match_all 有个旗子叫 PREG_OFFSET_CAPTURE 它将返回捕获的模式的偏移量，然后同一个函数具有可以设置的偏移量参数，所以我所做的是，我搜索案例1和案例2的模式，如果没有找到，则使用 PREG_OFFSET_CAPTURE ，然后进行第二次搜索，这次使用 PREG_OFFSET_CAPTURE 标记为偏移量，并返回行中唯一的日期。
希望这个解决方案对你有帮助。

<?php

    $subject = "22/10/2013 ånger, ska bet 6,428:-, 
    bet 6428, 2109+2016+2303 inlagt
    23/5/2014 bl 26/5, medd inga avgifter) 
     spelmissbruk 
     29/7/2014
    29/12/2014 föreslår godkänn
    27/4/2016 Föreslår godkänn
    CUSTOM TESTx 27/12/2018
    27/12/2018 - CUSTOM TEST 
    TESTVAL
    05/12/1987";
    $pattern = '- CUSTOM';
    $search = TRUE;

        /*Make sure you are searching a valid pattern*/
        if(!preg_match_all("/.*$pattern.*/",$subject, $matches)){
            echo "NO PATTERN MATCH";

        }
        /*Search the case that the date infront of the text*/
        if(preg_match_all("/(\d{1,2}\/\d{1,2}\/\d{4}).*$pattern.*/i",$subject, $matches))  {
            $date = $matches[0][0][0];

        }
        /*Search the case that the date is after the text*/
        if(preg_match_all("/.*$pattern.*(\d{1,2}.\/\d{1,2}\/\d{4})/",$subject, $matches))  {
            //echo "MATCH LAST";
            $date = $matches[0][0][0];
        }
        if(preg_match_all("/.*$pattern.*/",$subject, $matches,PREG_OFFSET_CAPTURE )){
            //echo " CAPTURED OFFSET ".$matches[0][0][1];
            if(preg_match_all("/(\d{1,2}.\/\d{1,2}\/\d{4})/",$subject, $dateMatches,PREG_OFFSET_CAPTURE,$matches[0][0][1]))  {
                //echo " CAPTURED OFFSET DATE";
                $date = $dateMatches[0][0][0];
            }

        }

    if(isset($date)){
        echo $date;
    }

    ?>