Sorry about your lose.
Here is a link about CMV and coronary artery disease that tells you some
problems related to infections and trying to determine their role in
disease,  it is not just infection but the host response to infection that
is important.

http://circ.ahajournals.org/cgi/content/full/100/13/1366

To your first question, I remember reading an editorial written by Dr. Ira
Shulman of USC, which was telling blood bankers to convey the presence of
cold agglutinins in an effective manner to the clinicians. There was a case
history mentioned in which a lymphoma diagnosis was delayed because even
though the blood bank found a direct coombs positive cold agglutinin, the
doctor was unable to interpret that reaction. I interpreted that to mean
that the pathologist should have consulted the doctor on that or a written
comment by the pathologist would been better.

With regard to the CBC, the morphology can be overlapping in reactive
lymphocytosis and lymphoma. Flow cytometry studies are needed to see the
monoclonality of any cells present denoting a lymphoma.
There is one reference of the presence of CMV infected endothelial cells
present in the feather tip of the smear where all the large cells found. It
is unusual to see large endothelial cells on smear but would suggest CMV
especially if nuclear inclusion are present with special stains.
CMV infected cells may also been seen in urine samples for cytology.

I remember the last case I was involved with was a gram stain the doctor
wanted my to review for pathogens looking for a cause of the pneumonia. It
looked like normal flora and I told her so.
She also indicated that their was liver involvement and it still did not
register. Viral pneumonia gives normal flora sputum gram stains and CMV
involves the liver. It is similar to EB mononucleosis. The patient succumbed
within a day, after the PCR results came back positive in most sites tested.