Platelet count decline and high neutrophil count within the first day of admission for painful sickle cell vaso-occlusive episodes predict severe complications.

Vaso-occlusive pain crisis (VOC) is the primary reason for emergency room (ER) visits and hospitalizations in sickle cell disease (SCD).1 Organ damage can occur in the lungs and other visceral organs during VOC.2-4 Platelet count decline was reported to predict acute chest syndrome5, 6 and other severe complications7 during hospitalization for VOC. We conducted a retrospective study in adult SCD patients admitted to University of Illinois Hospital over 1 year, to examine the trajectory of platelet counts and neutrophil counts8-10 before and during VOC admissions and the relationship of these changes to the development of severe complications. Electronic records of patients admitted to the inpatient sickle cell service from 7/1/2017 to 6/30/2018 were reviewed. Inclusion criteria were age at admission ≥18 years, diagnosis of SCD confirmed by haemoglobin fractionation or electrophoresis, and admission for management of a VOC with a pain score of 8/10 to 10/1011 at presentation. Exclusion criteria included haemoglobin S trait, elective admission for surgery, admission for accident or injury or admission for delivery of an infant. Complications recorded within each admission included: (1) acute pulmonary events defined by acute chest syndrome, new infiltrate on chest X-ray or use of ventilator; (2) acute kidney injury; (3) acute liver injury; (4) thrombotic or haemorrhagic stroke; (5) venous thromboembolism including deep vein thrombosis, atrial thrombosis and pulmonary embolism; (6) need for transfer to the intensive care unit (ICU); (7) need for exchange blood transfusion; and (8) death during hospitalization (Data S1). Data were initially collected for 559 admissions of 144 patients. The number of admissions per patient ranged from 1 to 36. Patients with frequent hospitalizations (>10) were less likely to have a complicated admission (binomial regression OR = 0.27, 95% confidence interval [CI] 0.15–0.48), suggesting their distinct aetiology of VOC. Excluding 12 heavy utilizers, 371 admissions remained for 132 patients, 41% having single admission. The patients had a median age of 35 years (range 20–74 years); 90 (68%) were female, 103 (78%) had severe sickle genotypes such as Hb SS or Sβ0-thalassemia, and 51 (39%) were on hydroxyurea (Table S1). As expected, complications overlapped considerably (Table S2). Over the course of the year studied, 27% admissions involving 45% unique patients were complicated (Table 1). Older age and severe sickle genotypes were associated with complicated admissions (Table S3). Complete blood count (CBC) was recorded for five time points: (1) Baseline, the most recent CBC determined at steady state (outpatient visit >3 weeks and <2.5 years before admission); (2) ER, at presentation to the ER or acute care centre; (3) First Morning, the first laboratory value on the day after admission; (4) Platelet Nadir, lowest platelet count during the admission before any exchange transfusion12; and (5) Discharge. To reflect change over time, proportional platelet or neutrophil counts with respect to Baseline (count ratio ER/Baseline, First morning/Baseline and Platelet Nadir/Baseline) and ER (First Morning/ER and Platelet Nadir/ER) were also assessed. Among the first admissions, median (interquartile range [IQR]) platelet counts at Baseline, ER, First Morning, Platelet Nadir and Discharge were 320 (239–388), 284 (220–369), 260 (192–329), 223 (174–312), and 272 (195–362) × 106/mL, respectively (Figure S1A). At baseline 3.8% of the platelet counts were below 150 × 106/mL and this increased to 8.3% at ER, 13% at First Morning and 18% at Platelet Nadir. Median (IQR) proportional decline from Baseline to Platelet Nadir was 0.24 (0.061–0.40) (Figure S1B). Integrating all admissions, average platelet counts declined from Baseline to Nadir by 76 × 106/mL; 22% of this decline occurred from Baseline to ER, 48% from ER to First Morning, and 30% from First Morning to Platelet Nadir (Figure 1A). Integrating all admissions, difference in platelet counts between complicated versus uncomplicated admissions was estimated adjusting for age, gender, sickle genotype severity and hydroxyurea treatment (Table S4). Platelet counts were 32 (95% CI 9.4–55) × 106/mL lower at First Morning and 36 (15–56) × 106/mL lower at Platelet Nadir in complicated versus uncomplicated admissions (Figure 1B). Proportional platelet counts relative to Baseline were 0.099 (95% CI 0.034–0.16) lower at First Morning and 0.13 (0.072–0.19) lower at Platelet Nadir in complicated versus uncomplicated admissions (Figure 1B). Proportional platelet counts relative to ER were also lower in complicated admissions (Figure 1B). Integrating all admissions, average neutrophil counts increased from Baseline to ER by 3.0 × 106/mL and then decreased progressively toward the baseline level but were still 0.76 × 106/mL above Baseline at Platelet Nadir (Figure 1C). Neutrophil counts were 1.3 (95% CI 0.20–2.4), 1.8 (0.57–3.1) and 2.6 (1.6–3.6) × 106/mL higher at ER, First Morning, and Platelet Nadir, respectively, in complicated versus uncomplicated admissions. Proportional neutrophil counts relative to Baseline were 0.28 (95% CI 0.017–0.55) higher at First Morning and 0.39 (0.19–0.58) higher at Platelet Nadir in complicated versus uncomplicated admissions (Figure 1D). Proportional neutrophil counts relative to ER were also higher in complicated admissions (Figure 1D). The temporal trend of low/decreasing platelet counts with severe complications tracked with that of high/increasing neutrophil counts (Spearman's ρ = −0.78, p = 0.017 across standardized β coefficients), implying a concordance of these two variables during the progression to a complicated hospitalization.13 Focusing on the first admissions without missing values (n = 130), a predictive score for complicated admission was developed using method similar to Bartolucci et al.14 Predictors included median-dichotomized age (>35 years), proportional platelet counts First Morning/Baseline (<0.8), and neutrophil counts First Morning (>7 × 106/mL), and severe sickle genotypes. Complicated admissions were modelled by logistic regression. The rounded coefficients of the four predictors at linear scale were all equal to 1, resulting in a predictive score ranged 0–4. The area under the receiver operating characteristic curve of the score was 0.74 (95% CI 0.66–0.83) (Figure S2). At a cut-off of ≥3, the predictive score achieved a sensitivity of 0.72, specificity of 0.70, positive predictive value of 0.48 and negative predictive value of 0.87. For SCD patients, a rise in the neutrophils and fall in the platelets are expected in admissions for VOC. In complicated admissions, however, these changes are more pronounced and are seen in the ER and more prominently on the first morning after admission. A 20% fall in platelet counts within the first 24–48 h after admission, compared to baseline, may be a warning sign of impending complications. Victor R. Gordeuk conceived study; Xu Zhang, Jin Han, Binal N. Shah and Victor R. Gordeuk designed study; Jin Han, Binal N. Shah and Victor R. Gordeuk collected data from electronic medical records; Xu Zhang designed the analysis and analysed data; Xu Zhang, Jin Han, Binal N. Shah, Santosh L. Saraf and Victor R. Gordeuk contributed materials and tools; and Xu Zhang and Victor R. Gordeuk wrote and revised the paper with all authors providing critical comments. The study was supported by Takeda Pharmaceuticals U.S.A., Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited, a Delaware corporation, with offices at 95 Hayden Avenue, Lexington, MA 02421, USA. The authors declare no competing financial interests. The study was approved by the Institutional Review Board of the University of Illinois at Chicago. Data S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.