Blood transfusions in prostatectomies

logistic regression
survival analysis
When patients have prostatectomies to treat prostate cancer, the surgery may cause blood loss requiring a blood transfusion. Blood transfusions may cause adverse reactions and affect the immune system, so this could affect the success of the surgery and whether the prostate cancer recurs. Use logistic regression or survival analysis to explore an observational dataset of patients.
Author

Alex Reinhart

Published

October 28, 2025

Data files
Data year

2011

Motivation

Prostate cancer is the most common type of malignant cancer in men. One common treatment option is radical prostatectomy: complete surgical removal of the prostate gland. Unfortunately, the cancer often returns after surgery—in about one quarter of patients, it will recur within five years—so surgeons are interested in understanding factors associated with recurrence and how they can be managed.

One possible factor is the use of blood transfusions during surgery. If the patient loses too much blood during surgery, they will be given a blood transfusion. Most blood transfusions use blood from unrelated donors that matches the recipient’s blood type, but because the blood comes from a different person, it contains different immune cells. It is possible that this suppresses or interferes with the recipient’s immune system, making it more likely the cancer can recur. It is also thought that older blood cells—that is, those that have been in storage for longer before being used in the transfusion—may cause more immune disruption and thus be worse for the patient.

This data comes from an observational study of patients who received blood transfusions during their prostatectomy. The data includes the number of units of blood transfused and the age of the blood used. Patients were tracked for five years to determine if their prostate cancer returned after the surgery, so it is possible to investigate how the age of the transfused blood is associated with recurrence.

Data

Each observation is one patient who had a prostatectomy at the Cleveland Clinic between July 6, 1998 and December 27, 2007, and who had a blood transfusion during their surgery. There are 316 total patients included in this sample.

Data preview

prostatectomy-storage.csv

Variable descriptions

Variable Description
RBC.Age.Group Red blood cell storage duration, grouped (1 = less than 13 days, 2 = 13-18 days, 3 = more than 18 days)
Median.RBC.Age Median age of red blood cells transfused (conversion of RBC.Age.Group to days)
Age Patient age (years)
AA 1 if patient is African-American, 0 otherwise
FamHx 1 if patient has family history of prostate cancer
PVol Prostate mass when removed (grams)
TVol Tumor volume (1 = low, 2 = medium, 3 = extensive)
T.Stage Clinical T category, a measure of the stage of cancer, estimated before surgery. 1 = stages T1-T2a, indicating tumor can’t be felt or seen in imaging, or is felt/seen in half of only one side of the prostate. 2 = stages T2b-T3, indicating tumor is felt/seen in more than half of one side, is in both sides, or has spread outside the prostate.
bGS Biopsy Gleason score, discretized. Based on microscopic examination of cells in a biopsy. Low scores means the cells look normal, high scores mean they look abnormal. 0 = score 0-6, 2 = score 7, 3 = score 8-10.
BN+ Bladder neck positive: whether the neck of the bladder, at the edge of the prostate, contained cancerous cells when investigated after surgery. 1 = yes, 0 = no.
OrganConfined 1 if cancer is limited to the prostate gland, 0 if it has spread outside it
PreopPSA Prostate-specific antigen (PSA) levels measured before the surgery; levels are typically higher in people with prostate cancer (ng/mL)
PreopTherapy Whether the patient received cancer treatment prior to surgery (1 = yes, 0 = no)
Units Number of allogeneic units of blood received
sGS Surgical Gleason score, discretized. Same as bGS, but based on examination of the tumor removed in surgery, rather than on cells removed in a biopsy
AnyAdjTherapy Whether the patient received adjuvant therapy after the surgery, i.e., another treatment to prevent recurrence (such as chemotherapy, radiation, or other cancer treatments)
AdjRadTherapy Whether the patient received radiation therapy after surgery to prevent recurrence
Recurrence Whether the prostate cancer recurred (1 = yes, 0 = no)
Censor Whether the study ended before the patient had any recurrence, so their final status is unknown (1 = censored, 0 = not censored)
TimeToRecurrence Time to recurrence of prostate cancer, months

Questions

  1. It’s possible that the amount of blood transfused depends on features of the tumor—perhaps more advanced cases required more invasive surgery with greater blood loss, for instance. Conduct an exploratory data analysis to judge the relationship between the units of blood transfused and indicators of the severity of the tumor, such as clinical T category, Gleason score, and tumor volume.
  2. Explore how tumor severity is related to the risk of recurrence. Conduct an exploratory analysis and then use regression to predict the risk of recurrence. How much more likely is recurrence for patients with a tumor that is advanced at the time of surgery, as measured by clinical T category or surgical Gleason score?
  3. Analyze the main research question: Is the risk of recurrence related to the age of blood transfused? Is there evidence that older blood is associated with a higher risk of recurrence?

References

Dataset adapted for classroom use by: Amy S. Nowacki, “Blood Storage Dataset”, TSHS Resources Portal (2017). Available at https://www.causeweb.org/tshs/blood-storage/. Released under a CC BY-NC-SA license.

Original study: Cata et al. (2011). “Blood Storage Duration and Biochemical Recurrence of Cancer After Radical Prostatectomy.” Mayo Clinic Proceedings 86 (2), 120-127. https://doi.org/10.4065/mcp.2010.0313