I agree with tedrek++. The code you show looks fine for syntax.

A few suggestions, however. Your line 135, the first shown, is true unless @results is empty. The usual way to write that is

if (@results) {
    #...
}
[download]

That works because @results is in scalar context there, giving the number of elements. This decision is not strictly necessary, since the for loop will iterate zero times for an empty @results.

The evident duplication of code in your long run of decisions suggests condensing that section into a loop. A hash for the data with the printed terms "Title".. as keys seems inviting, but suffers from not having a canonical order. Let's just impose that order by defining an array of those terms,

my @columns = qw/ Title Author Year Journal
                 # ...
                  /;
for my $record (@results) {
    my $index = 0;
    for (split /\|/, $record) {
        printf "%s: %s$/", $columns[$index], $_ if $_;
        $index++;
    }
}
[download]

That condensation separates data from logic, making the whole easier to maintain. It's not the last word, but a step in the right direction.

Have you looked at any of the CSV modules?