We evaluated all consecutive patients ≥ 18 years with newly diagnosed acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML) or acute promyelocytic leukemia (APL) by the WHO criteria19 from January 2015 to March 2020, in the Hematology Department of the Instituto Nacional de Ciencias Médicas y Nutrición in Mexico City. Patients treated at the same Institution with secondary leukemia caused by the primary disease, who refused any treatment or did not have a follow-up, were excluded.

This retrospective cohort study included data extracted from the electronic medical record, baseline characteristics, such as age, with the adolescent and young adult (AYA) being < 39 years, gender, body mass index (BMI), socioeconomic level classification (low versus not low),20 education, Charlson comorbidity index (CCI), Eastern Cooperative Oncology Group (ECOG) performance status (0–2 versus > 2), presence of comorbidities, complete blood count (CBC), blast count in bone marrow aspiration, liver function tests (LFTs), creatinine level (mg/dl), presence of symptoms, fever, infection, tumor lysis syndrome (TLS), according to the Cairo-Bishop criteria, and number of consultations after the first symptoms were recorded. The AML was classified as low-, intermediate-, or high-risk, according to the European Leukemia Net.4 The APL was classified as low-, intermediate-, or high-risk as per the Sanz criteria.5 The ALL was classified as high-risk, according to the presence of at least one of the following criteria: hyperleukocytosis (white blood cells 30 × 109/L for B-cell ALL or > 100 × 109/L for T-cell ALL), or a complex karyotype.6 The chemotherapy was classified as intensive (Hyper-CVAD,21 CALGB 10,40322, 7 + 3 (7-day continuous infusion of cytarabine 100 to 200 mg/m2 associated with an anthracycline on days 1 to 3) and AIDA regimens23), non-intensive (mini-Hyper-CVAD without the antibody-drug conjugate therapy,24 low-dose cytarabine (20 mg subcutaneous twice per day on days 1 to 10), azacytidine monotherapy (75 mg/m2 subcutaneous per day on days 1 to 7)) or palliative (support treatment only)), according to the medical decision and available institutional regimens. The support treatment included prophylactic antibiotics, according to the ASCO/IDSA guidelines25 and the use of granulocyte colony-stimulating factor during neutropenic fever after the induction scheme, according to the treating physician. We defined the following intervals (number of days): patient interval (PI), from the first symptom to the first medical consultation (including symptomatic patients only); diagnostic interval (DI), from the first medical consultation to the diagnosis (the day of the bone marrow aspiration); treatment interval (TI), from the diagnosis to treatment initiation (including patients who received chemotherapy, or in APL patients, the day of the administration of the all-trans retinoic acid); total time interval (TTI), from the first symptom to the treatment initiation. We evaluated the complete response (CR) after the induction treatment and estimated overall survival (OS) to define the response to the treatment. The CR was established in accordance with the Cheson criteria, that states the absence of extramedullary leukemia, the lack of peripheral blood blasts, a bone marrow blast percentage below 5%, a neutrophil count ≥ 1.5 × 109/L, and platelets ≥ 100 × 109/L.26 The OS was the time interval between the diagnosis and the patient's death or the last day of the follow-up. Interval delays refer to periods over the interquartile range (IQR) of the median time expected. The ED was defined as that which occurred within 60 days of the diagnosis. We divided the ED into 30-day mortality and 60-day mortality to analyze the associated factors because both periods have been previously reported in the literature.

Standard descriptive statistics were expressed as median and IQR or mean and standard deviation, according to the normality, or, when appropriate, absolute counts and percentages. Differential characteristics of groups of patients according to the leukemia type and interval delay were compared by statistical tests (the χ2 test for categorical variables, Student's t-test and the ANOVA test for parametric variables and the Mann–Whitney U test and Kruskal–Wallis test for non-parametric variables). We considered a p-value < 0.05 statistically significant. The Kaplan–Meier approach was chosen for the OS, using the log-rank test. We calculated the odds ratio (OR) and performed a logistic regression to identify factors associated with an increased risk for an ED, including variables that were statistically significant (p < 0.05) in the univariate analysis. All analyses were performed with the SPSS software (version 22).

A total of 188 patients with newly diagnosed AL met the eligibility criteria and 96 (51%) of these were women, the median age was 37 years (IQR, 24–54) and 54.8% were AYA. The median BMI was 25.2 (IQR, 22.7–29.3) and 19.7% were obese (BMI > 30); 71 patients (37.8%) had at least one comorbidity; 12.2% had diabetes mellitus, 60.4% had a CCI < 3, and 95.7% (180 patients) had an ECOG of 0–2. Only 44.1% (83 patients) were residents of Mexico City; 20% (39 patients) had a university degree and 80.3% had a low socioeconomic level (LSL). Related symptoms were present in 97.3% of the patients. The patients consulted with a median of three physicians (IQR, 2–4) before consulting with a hematologist. The first medical contact was a primary care physician (63.2%) and the remaining went to the emergency room. There was no difference in the previous variables compared between AL subtypes.

The more prevalent AL was ALL in 102 patients (54.3%), followed by AML, 57 patients (30.3%) and APL, 29 patients (15.4%). According to the previously mentioned criteria, of 184 patients, 61.2% were high-risk, 31.4%, standard/intermediate-risk and 5.3%, favorable-risk; 4 patients had no risk classification due to incomplete data. The blood test, clinical characteristics and the median interval times of the ALs are shown in Table 1 and Figure 1.

Figure 1.

Time intervals (days) of acute leukemias.

(0.09MB).

PI: Patient interval; DI: Diagnostic interval; TI: Treatment interval; TTI: Total interval; ALL: acute lymphoblastic leukemia; AML: acute myeloblastic leukemia; APL: acute promyelocytic leukemia.

Intensive induction therapy, according to the institutional protocol, was offered to 164 patients (87.2%), non-intensive induction therapy to 13 patients (6.9%) and palliative care to 11 patients (5.9%) (Table 2). After the diagnosis, the median hospital stay was 34 days (IQR, 26–44). The induction complications: pneumonia in 42%, blood-stream infection, 38%, septic shock, 27%, acute complicated urinary tract infection, 23%, sinusitis, 17%, soft tissue infection (STI), 16%, Clostridioides difficile infection, 11%, neutropenic enterocolitis, 8%, odontogenic infection, 7%; hepatotoxicity 35% (G3 - 4 in 37 patients), bleeding, 33% (G3 - 4 in 31 patients), acute kidney injury, 22% (G3 in 13 patients), ischemic stroke, 1.5%, and; appendicitis, 5%. Forty-eight patients (25.5%) required intensive care and 34 (18%), invasive mechanical ventilation.

The gram-negative bacilli (42.6%) and gram-positive cocci (8.5%) were identified in culture isolates. In total, 57 patients were suspected of having a fungal infection and, among these, 22 (12.5%) had a proven fungal infection. The previous complications that were different among the AL types are shown in Table 2.

The CR was achieved in 76.6% of the AL patients. An ED occurred in 39 patients (20.7%). The main causes of the ED were infection in 25 patients (64%), bleeding in 5 (12.8%) and other causes in 9 (23%). Most of them (66%) died in the first 30 days. At the time of the analysis, 51.5% of the cases had survived (ALL, 47.1%, AML, 43.9%, and APL, 82.8%; p = 0.001), with a median follow-up of 12.9 months (IQR, 3.86 – 25.6) (Table 2). The median OS of each type of AL is shown in the Figure 2, with ALL at 19.5 months (95% CI 14.5–24.6), AML, 14 months (95% CI 2–26), and APL not reached.

Figure 2.

Overall survival between AL types. Kaplan–Meir curve.

(0.13MB).

ALL: acute lymphoblastic leukemia; AML: acute myeloblastic leukemia; APL: acute promyelocytic leukemia.

All the variables associated with an ED are shown in Table 3. In patients with an ED (n = 39, 20.7%), none of the intervals were statistically significant, compared to those who survived (n = 149, 79.3%). Among patients with an ED in the first 30 days (n = 26, 13.8%), compared to those who survived (n = 162, 86.2%), the PI was 7 days (IQR, 2–13) vs. 15 days (IQR, 7–28; p = 0.004), respectively.

Forty-six patients (24.4%) met the criteria for the TTI delay (> 52 days). When compared to patients without the TTI delay criteria (n = 142, 75.6%), we found that the main associated factors in the univariate analysis that prevented the TTI delay were the first medical contact at the emergency room (p = 0.019) and the presence of fever at diagnosis (p = 0.042).

Thirty-nine patients (21.3%) had a PI delay (> 27 days), compared to those who did not (n = 149, 78.7%). An infection at diagnosis (p = 0.041), ICU at admission (p = 0.041) and a lower first 30-day mortality (p = 0.022) were statistically significant protective factors. Furthermore, patients with a PI delay had a longer DI (15 vs. 9 days; p = 0.012), lower hemoglobin levels (7.26 vs. 8.6 g/dl; p = 0.039), higher albumin levels (4.1 vs. 3.9 g/dl; p = 0.015), and a longer hospital stay (38 vs. 33 days; p = 0.039).

Forty-two patients (22.3%) had a DI delay (> 20 days), compared to those who did not (n = 146, 87.3%). The first medical contact at the emergency room (p = 0.024), neutropenic colitis (p = 0.032) and the STI (p = 0.001) were statistically significant protective factors. Furthermore, patients with a delayed DI had a lower BMI (24.1 vs. 25.9; p = 0.038), lower creatinine levels (0.74 vs. 0.86 mg/dl; p = 0.006) and longer TI (5 vs. 3 days; p = 0.002), compared to those who did not.

Twenty-five patients (14.1%) had a TI delay (> 6 days). None of the patients with APL were included in the group with the TI delay. Up to 46.2% of the patients who received a low-intensity chemotherapy scheme were in the TI delay group (p = 0.001). Patients with a TI delay had an increased association with a higher ECOG > 2 (p = 0.008) and septic shock (p = 0.014), compared to those who did not (n = 163, 86.9%). Furthermore, patients with a TI delay had lower hemoglobin levels (7.05 vs. 8.4; p = 0.012), lower LDH levels (330 vs. 565 UI/L; p = 0.026), a higher number of previous consultations (4 vs. 3; p = 0.025) and a shorter follow-up (227 vs. 442 days; p = 0.036), compared to those without a TI delay. When the OS of patients with and without a TI delay was analyzed by the Kaplan-Meier method, it was not statistically significant (log rank, p = 0.136).

None of the interval delays were associated with an ED.