< Back to previous page


Bacterial contamination of blood products for transfusion in the Democratic Republic of the Congo: prevalence, risk analysis, prevention

There are limited data on bacterial contamination of blood products for transfusion in sub-Saharan Africa. The available data vary largely over the studied countries, with contamination ratios ranging from 0.0 – 17.5%. Skin and environmental bacteria are the main sources of blood product contamination. Gram-negative bacteria specifically can cause severe transfusion reactions. The objectives of the present studies were (i) to assess the prevalence of bacterial contamination of blood products in the Democratic Republic of the Congo (DRC), (ii) to assess presence of Gram-negative bacteria on blood donor skin and in the blood bank environment, to determine associated bacterial counts, and to evaluate skin antisepsis, (iii) to assess knowledge, attitude and practices concerning bacterial contamination of blood products among healthcare workers .

In Chapter 1, blood products in one rural and two urban hospitals (paediatric and general) in DRC were cultured in a blood culture bottle (4mL) and on an agar-coated slide to estimate colony forming units (CFU/ml). Bacteria were identified with biochemical tests and MALDI-TOF. Exposure time > 10°C was assessed on a subset of blood products. We found that 1.4% blood products (whole blood and red cell concentrates) were contaminated with bacteria, at counts ≤ 10³ CFU/ml. Findings were lower compared to most other studies in sub-Saharan Africa, but comparable to cross-sectional results from high-income countries. As expected, cold chain maintenance was challenging in all study sites, with all blood bags being exposed to temperatures >10°C for extended time. Contamination ratio was lowest in the rural hospital and highest in one urban hospital, where the only Gram-negative bacteria were isolated. Study site was the only statistically significant associated factor. We did not demonstrate significant differences among donor type, season, type and age of blood product, and blood exposure time >10°C.

In Chapter 2, among blood donors at the National Blood Transfusion Centre (NBTC) and at a rural hospital in DRC, the antecubital fossa skin of the non-disinfected arm (not used for blood collection) was swabbed (25cm² surface) and cultured for total and Gram-negative bacterial counts. Bacteria were identified with MALDI-TOF. For evaluation of the NBTC antisepsis procedure (i.e. ethanol 70%) culture results of the disinfected arm (used for blood collection) were compared with those of the non-disinfected arm.  We demonstrated presence of clinically relevant Gram-negative bacteria on their arm skin. Gram-negative bacteria were isolated for 43.8% of sampled donors, Staphylococcus aureus from 3.8% of donors. Non-fermentative Gram-negative rods predominated among species, incl. species previously described in transfusion reactions (e.g. Pseudomonas spp., Stenotrophomonas, Enterobacter, Acinetobacter, …). Skin antisepsis with ethanol 70% according to NBTC standard procedures did not fulfil study-defined acceptance criteria of ≤ 2 CFU/25cm² after antisepsis. Gram-negative bacteria were still isolated among 5.6% of blood donor participants post-antisepsis.

In Chapter 3, fomite surfaces, liquids, and air in DRC blood banks were sampled. Samples were cultured by 25 cm² contact plates (fomite surfaces), plate-streak method (liquids), membrane filtration (water), and 500L active air sampling on selective agars. After 48h incubation at 35°C, colonies were counted, Gram-negative species were identified by MALDI-TOF. Gram-negative bacteria were isolated from multiple sources: most water and air samples and half of fomite surfaces and soaps. Part of the isolated bacterial species had been previously described in (fatal) transfusion reactions (Enterobacter cloacae, Acinetobacter baumannii/nosocomialis, Klebsiella pneumoniae, Serratia marcescens, …). Bacterial counts were high, with most samples exceeding guideline thresholds.  High touch surfaces, liquid soap samples and water were most at risk to contain pathogenic Enterobacterales at high counts.

In Chapter 4, in three hospitals and the NBTC, two multiple-choice knowledge-attitude-practice surveys were completed on a tablet computer: one each, for blood bank (31 questions) and for clinical ward staff (20 questions). A score was calculated for 11 overlapping knowledge questions. The survey results showed that healthcare workers were insufficiently aware of risk factors and associated prevention measures of bacterial contamination of blood products. Blood bank staff scored higher than clinical ward staff, doctors scored higher than nurses. Presence of bacterial contamination in the hospital environment and on blood donor skin was underestimated, septic transfusion reactions were not well recognized and challenges specific to low-resource settings were identified (power failure challenges cold chain, blood products transported without cool box by patient attendants, recycling of single-use containers for soap and disinfectants).

In conclusion, although contamination ratios of blood products were lower than expected, our findings still correspond to two transfused contaminated blood products per week in the studied hospitals. Regarding already available preventive measures of bacterial contamination (donor selection, skin antisepsis, cold chain maintenance), their importance needs to be emphasized to healthcare workers and their implementation improved. Additional low-tech and low-cost measures are the use of diversion pouches, minimizing the open blood system, and environmental cleaning. Implementation of new measures or improvement of existing measures needs to be accompanied by hands-on training, monitoring and feedback.

Date:1 Jan 2017 →  19 Dec 2023
Keywords:transfusion, bacteria, contamination
Disciplines:Microbiology, Systems biology, Laboratory medicine
Project type:PhD project