The aim of the study is to evaluate the effects of red blood cell (RBC) transfusions on pulmonary parameters in critically ill, non-bleeding patients. Retrospective chart analysis was performed on critically ill patients without overt bleeding in the intensive care unit (ICU) of a university hospital. In 83 patients in a 5-month period, who had received at least 1 RBC unit and stayed at least 24 h in the ICU, 199 transfusions of median 2 RBCs per transfusion (n = 504) were studied. Pulmonary parameters were retrieved during the period between 24 h before the start of transfusion and 24–48 h after transfusion. Outcome was assessed. The PaO2/FIO2 dose-dependently decreased from 250 ± 105 at baseline to 240 ± 102 mmHg at 24 h after RBC transfusion (P = 0·003), irrespective of acute lung injury at baseline and RBC storage time. The lung injury score (LIS) also increased dose-dependently, whereas, at 48 h, oxygenation and LIS largely returned to baseline. For every seven RBCs transfused, the LIS transiently increased by 1 unit. There were no changes in haemodynamics, lung mechanics or chest radiography. The total number of RBCs given in the ICU did not directly contribute to ICU and 1-year mortality prediction. Transfusion of RBCs decreases oxygenation thereby increasing the LIS, dose-dependently and transiently, in a heterogeneous population of critically ill, non-bleeding patients, independent of prior cardiorespiratory status and RBC storage time. The effects are subtle, may go unseen and unreported and may represent subclinical transfusion-related acute lung injury. They do not adversely affect outcome, even at 1-year follow-up.