Background Complete blood count derived indexes such as lymphocyte-to-neutrophil ratio (NLR) may help in predicting pneumonia and prognosis in acute stroke. However, the optimal time point for using these biomarkers is not known. Methods In 205 consecutive severe (NIHSS>10) acute ischemic stroke patients, daily leukocyte, lymphocyte, neutrophil, monocyte, platelet, albumin, fibrinogen, hematocrit, NLR, PLR (Platelet-to-lymphocyte-ratio), LMR (Lymphocyte-to-monocyte-ratio), and SII (systemic-immune-inflammation-index) were determined. General linear models for repeated measures (GLMR) and receiver operating characteristics [ROC] analyses were conducted to define their daily discriminative ability. Results GLMR-prognosis modeling documented that the main determinants of significant daily variations of 12 parameters studied were age and 24th-hour-NIHSS. In addition, daily changes of NLR, neutrophil, leukocyte (all increased on day-2 and remained higher) and platelet count (decreased after day-6 and stayed lower) were related significantly to survival status (mortality in 19.5%). Albumin levels (lower after day-2) were marginally associated by functional prognosis (modified-Rankin-Score <= 3 in 28%). There was a borderline relationship (p = 0.05) between NLR (between day-1 and day-8) and pneumonia development (in 36%). Useful discrimination capability (95% confidence interval lower limit of area-under-curve of ROC >= 0.7) was noted for NLR measured on day-6 for mortality, NLR (for 6 days, from day-3-to-day-7, and day-11) and albumin (for every day except day-11 after day-4) for reasonable prognosis and none for pneumonia development. Conclusions Inflammatory parameters from peripheral routine blood tests showed significant variations during the first two weeks following stroke, but discriminative capacity of these changes is limited due to confounders such as age and post-treatment clinical stroke severity.