BACKGROUND: The problem of spurious hyperkalaemia in blood samples delivered to hospitals from local general practitioners' (GP) surgeries remains a source of diagnostic confusion and potential danger to patients. We have carried out retrospective and prospective audits of serum potassium measurements to assess the influences of blood sample delivery time and temperature during transit from GP surgeries to a centralized laboratory on the risk of spurious serum potassium measurements. METHOD: The retrospective audit included serum potassium measurements made by the pathology laboratory at Addenbrooke's Hospital on 51 843 patients attending either the hospital itself or one of 62 local GP surgeries during the months August 2001 to February 2002 (i.e. summer through to winter). The probability of spurious potassium measurements on GP patients relative to blood sample delivery time and temperature was modelled by least-squares multiple linear regression. The prospective audit included serum potassium measurements made on 40 patients presenting early in the day at the Sutton GP surgery near Ely, Cambridgeshire, during the months February to May 2002. RESULTS: Low blood sample delivery temperature had a greater impact on the probability of spurious potassium measurements than long delivery time. Low delivery temperature also had a greater impact on the probability of spurious hyperkalaemia than it had on the probability of spurious normokalaemia. At temperatures below 20.3 degrees C, the probabilities of both spurious hyper- and normokalaemia for samples delivered from the most distant and closest GP surgeries to Addenbrooke's increased significantly. CONCLUSIONS: Maintaining blood sample delivery temperatures a few degrees above 20 degrees C will minimize the risk of spurious serum potassium measurements on samples from GP patients.
BACKGROUND: The problem of spurious hyperkalaemia in blood samples delivered to hospitals from local general practitioners' (GP) surgeries remains a source of diagnostic confusion and potential danger to patients. We have carried out retrospective and prospective audits of serum potassium measurements to assess the influences of blood sample delivery time and temperature during transit from GP surgeries to a centralized laboratory on the risk of spurious serum potassium measurements. METHOD: The retrospective audit included serum potassium measurements made by the pathology laboratory at Addenbrooke's Hospital on 51 843 patients attending either the hospital itself or one of 62 local GP surgeries during the months August 2001 to February 2002 (i.e. summer through to winter). The probability of spurious potassium measurements on GP patients relative to blood sample delivery time and temperature was modelled by least-squares multiple linear regression. The prospective audit included serum potassium measurements made on 40 patients presenting early in the day at the Sutton GP surgery near Ely, Cambridgeshire, during the months February to May 2002. RESULTS: Low blood sample delivery temperature had a greater impact on the probability of spurious potassium measurements than long delivery time. Low delivery temperature also had a greater impact on the probability of spurious hyperkalaemia than it had on the probability of spurious normokalaemia. At temperatures below 20.3 degrees C, the probabilities of both spurious hyper- and normokalaemia for samples delivered from the most distant and closest GP surgeries to Addenbrooke's increased significantly. CONCLUSIONS: Maintaining blood sample delivery temperatures a few degrees above 20 degrees C will minimize the risk of spurious serum potassium measurements on samples from GP patients.