A total of 212 patients with de novo APL were hospitalized at Nanfang Hospital of Southern Medical University (Guangzhou, China) between January 2003 and December 2014. Of these patients, 169 accepted ATRA and cytotoxic agents during induction treatment, among which 96 were categorized into a low-intermediate risk group, and the remaining patients categorized into a high risk group, based on Sanz's risk stratification model for survival prediction (14). All 96 patients [47 males and 49 females with a median age of 32 years (range 15–66)] with low-intermediate risk were retrospectively enrolled in the present study. Disease diagnosis was confirmed by bone marrow aspiration, chromosome karyotyping analysis, fluorescence in situ hybridization analysis and polymerase chain reaction tests. Informed consent prior to and regarding the treatment protocol was obtained from all patients analyzed in the present study.

When a diagnosis of APL was suspected, ATRA (30 mg/m²/day) was administered as early as possible, until CR was achieved. Chemotherapy comprised treatment with idarubicin (8 mg/m²/day on days 1, 3 and 5), daunorubicin (45 mg/m²/day on days 1, 3 and 5) or homoharringtonine (2 mg/m²/day on days 1–5); induction treatment was combined with cytarabine (100 mg/m²/day on days 1–7) and/or hydroxyurea (2.0–3.0 g/day, adjusted according to regular blood tests). In addition, 36 patients simultaneously received ATO (0.15 mg/kg/day for 14 days). Chemotherapy was applied 0–21 days after initiation of ATRA (median time, 7 days; range, 0–20 days). Blood product support was applied to maintain a platelet (PLT) level ≥30×10⁹/l, hemoglobin level ≥70 g/l and plasma fibrinogen level ≥1.5 g/l. All adverse events related to treatments, including bone marrow depression, infection and bleeding were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.03 (15).

Retinoic acid syndrome, also known as DS, was diagnosed based on the incidence of at least two of the following clinical features: unexplained fever, acute respiratory distress with interstitial pulmonary infiltrates, acute renal failure, weight gain >5 kg, unexplained hypotension and pleuropericardial effusion (13). Prevention strategies included dexamethasone (10 mg q12h) and discontinuation of ATRA and ATO. All 96 patients were categorized according to Sanz's risk stratification score (14) as low risk (WBC count <10.0×10⁹/l and PLT ≥40.0×10⁹/l at diagnosis) and intermediate risk (WBC count <10×10⁹/l and PLT <40.0×10⁹/l). CR was defined as the presence of <5% blast cells in bone marrow aspirates, PLT >100×10⁹/l and no juvenile cells in peripheral blood according to the criteria set by the US National Cancer Institute (16). Early mortality was defined as fatality during induction treatment from the first day of hospitalizations. Eastern Cooperative Oncology Group performance status score (0–4 score) was applied to investigate patient's physical status prior to chemotherapy (17).

In most cases, WBC count increased as ATRA was used. Thresholds were initially set according to the WBC count at chemotherapy initiation as ≤4×10⁹/l, >4×10⁹ and <10×10⁹/l, and ≥10×10⁹/l, since normal WBC count is between 4×10⁹ and 10×10⁹/l. However, no significant difference was determined using these classes. The thresholds were therefore altered until it was apparent when use of cytotoxic agents was appropriate. Thus, based on the WBC count at chemotherapy initiation, the 96 patients were divided into three groups: low WBC group (WBC count ≤4×10⁹/l), intermediate WBC group (WBC count >4×10⁹ and <15×10⁹/l), high WBC group (WBC count ≥15×10⁹/l). Table I presents a comparison of the baseline clinical and laboratory parameters of the low, intermediate and high WBC groups. According to the period from ATRA commencement to chemotherapy, the 96 patients were also divided into two groups: ≤3 days group (treated with chemotherapy within 3 days of ATRA) and >3 days group (treated with chemotherapy at least 3 days after ATRA). Table II presents a comparison of the baseline clinical and laboratory parameters of these two groups.

Statistical analysis was performed with SPSS v.17.0 software (SPSS, Inc., Chicago, IL, USA). All data were collected in January, 2015. Clinical features are presented as percentages (%) for categorical variables and as mean values ± standard deviation for normally distributed continuous variables. The χ² test was used to analyze the significance of differences in the distribution of categorical variables between the patient subsets, and Bonferroni's correction was used post-hoc to compare every two groups. The unpaired Student's test or Mann-Whitney test was used to analyze the significance of differences in the distribution of continuous parametric variables and the distribution of ranked variables. Multivariate analysis was performed by using a binary logistic regression model. WBC count at diagnosis, PLT count at diagnosis, WBC count at chemotherapy initiation and the period from ATRA to chemotherapy were considered in the multivariate analysis to evaluate their effects on DS, early mortality and remission. P<0.05 was considered to indicate a statistically significant difference.

Comparison of the baseline clinical and laboratory parameters showed no notable differences between these groups (Table I). As depicted in Table III, the incidence of DS was 0.0, 11.1 and 40.0% respectively in the low, intermediate and high WBC groups, respectively (P<0.001). DS incidence was significantly lower in the low and intermediate WBC groups compared with in the high WBC group (P=0.001 and 0.003). There was no difference in the extent of bone marrow suppression between the three groups. Grade 3–4 infection rates were 71.4, 33.3 and 43.3% (P=0.015). Multiple comparisons indicated grade 3–4 infection in the low WBC group was significantly more prevalent than that in the intermediate WBC group (P=0.004). Grade 3–4 bleeding rates were 9.5, 26.7 and 40.0%, respectively (P=0.055), which in the low WBC group was significantly lower than in the high WBC group (P=0.016). CR rates were 90.5, 100 and 73.3%, respectively (P<0.001), which of the intermediate WBC group was significantly higher than of the high WBC group (P=0.001). Early mortality rate in each group was 4.8, 0.0 and 26.7%, respectively (P<0.001), and determined as significantly higher in the high WBC group than in the intermediate WBC group (P=0.003). The time to achieve CR was similar between the three groups (P=0.498; Table III).

Comparison of the baseline clinical and laboratory parameters revealed that PLT count was higher in the >3 days group (P=0.001). Time between ATRA and chemotherapy was also associated with the distribution of Sanz's risk stratification (P=0.012; Table II). Regarding the initiation time of chemotherapy, there were no significant differences in DS incidence, grade 3–4 bone marrow suppression rate, grade 3–4 infection rate, CR rate and early mortality rate between the ≤3 days and >3 days groups (Table IV). Time to achieve CR was longer in the >3 days group (37.6±10.0 vs. 30.8±11.7 days; P=0.004). Although PLT count was higher in the >3 days group, bleeding related to chemotherapy was also more prevalent in the >3 days group (33.3 vs. 7.4%, P=0.009; Table IV). To eliminate the effect of PLT count, multivariate analysis was performed to correct for bias.

The results of multivariate analysis revealed that WBC count at chemotherapy initiation was an independent risk factor for the occurrence of DS [P=0.002, odds ratio (OR) =1.058, 95% confidence interval (CI) =1.021–1.095]. WBC count at chemotherapy initiation was also indicated to influence the occurrence of early mortality (P=0.036, OR =1.036, 95% CI =1.002–1.070). However, the period from ATRA to chemotherapy, WBC count at diagnosis and PLT count at diagnosis were not independent risk factors for DS, early mortality or remission failure (Table V).