Tumor Lysis Syndrome Is Associated with Worse Outcomes in Adult Patients with Acute Lymphoblastic Leukemia Free

Tumor lysis syndrome (TLS) is a common complication in patients with hematological malignancies that can occur in response to cytotoxic treatment or, spontaneously, that has been typically associated with acute leukemia and high-grade lymphomas. It is a metabolic emergency that develops after tumor cells lyse and release their contents into the circulation and is characterized by hyperkalemia, hypophosphatemia, hyperuricemia, and/or hypocalcemia. These abnormalities sometimes lead to end-organ damage such as acute kidney injury (AKI), cardiac arrhythmias, seizures, and/or death [1, 2].

In 2004, Cairo and Bishop coined the definitions of TLS, and Howard et al. updated them in 2011. Laboratory TLS (L-TLS) was defined as two or more of the following abnormalities 3 days before or up to 7 days after the initiation of cytotoxic therapy: uric acid greater than 8 mg/dL, phosphorus greater than 4.5 mg/dL, potassium greater than 6 mmol/L, and corrected calcium less than 7 mg/dL or ionized calcium less than 4.5 mg/dL. Clinical TLS (C-TLS) was defined as L-TLS accompanied by end-organ damage caused probably or definitely by these laboratory abnormalities: AKI (an increase of at least 0.3 mg/dL in the creatinine or an average urine output of less than 0.5 mL/kg/h for at least 6 h), cardiac arrhythmia, seizure, or sudden death. The revision of the Cairo-Bishop definition also considered symptomatic hypocalcemia as part of the spectrum of C-TLS, and they eliminated 25% change from the baseline of the L-TLS definition as they stipulated that these increases or decreases are not clinically relevant [1, 3].

An expert panel consensus stratified the risk of malignancy to develop TLS into low, intermediate, or high based on the disease-specific biology, therapy to receive, lactate dehydrogenase (LDH) value and white blood cell (WBC) count, presence of renal dysfunction or an isolated elevation of uric acid, phosphate, or potassium, and prophylaxis and monitoring based on this stratification. Adults with acute lymphoblastic leukemia (ALL) only fit into intermediate- or high-risk disease, the latter defined as the presence of an elevated LDH 2 times greater than its upper laboratory normal (ULN) value and/or a WBC count above 100×10⁹/L, with the intermediate-risk disease having none of the mentioned [4].

Epidemiological data about TLS in adults with ALL are scarce. From a total of 28,370 hospital admissions for TLS from 2010 to 2013 in the USA, 13% were because of ALL [5]. A retrospective multicenter survival analysis of 559 adults with ALL from Mexico City described a prevalence of TLS of 9.8% [6], and a prospective cohort study of 153 patients with hematological malignancies showed TLS occurrence of 30.7%; 3 of the 6 patients with ALL that were included developed this complication [7]. Trials evaluating treatment in ALL rarely report TLS incidence. The CALGB 10403 trial reported a TLS incidence of 5% [8].

The development of TLS-related AKI or arrhythmia has been associated with higher mortality and longer hospital stay [9‒11], but risk factors for developing this complication and its outcomes in adults with ALL have not been described yet. Effort has been made in adults with acute myeloid leukemia, and some TLS risk-prediction models have been published [12‒14]. Thus, to characterize the epidemiology of TLS and to identify factors associated with the development of this complication and its outcomes in adults with ALL, we conducted a retrospective case-cohort study in a tertiary care center in Mexico City, Mexico.

This is a retrospective case-cohort study that included 138 patients, aged 18 or more, with a newly diagnosed ALL according to the WHO criteria [15], admitted to our institution, Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, a tertiary care center in Mexico City, who were treated with their first induction chemotherapy regimen between January 2015 and October 2022. We excluded patients receiving only supportive/palliative care. Clinical data and laboratory charts during induction were reviewed.

TLS (L-TLS and C-TLS) and their corresponding laboratory cut-off values were defined following the previously mentioned definitions [1, 3]. The LDH ULN value in our institution is 271 U/L. High-risk ALL was defined as patients with at least one of the following characteristics: age greater than 35 years, WBC count above 30 × 10⁹/L for B-ALL or 100 × 10⁹/L for T-ALL, or high-risk cytogenetics (BCR-ABL1 philadelphia chromosome-positive, MLL rearrangements, hypodiploidy [30–39 chromosomes], or complex karyotype [5 or more chromosomal abnormalities]) [16]. ALL with a high risk for TLS was defined based on the expert consensus mentioned above [4]. Complete response (CR) was defined as less than 5% of blasts on bone marrow after induction regimen, measurable residual disease was tested in bone marrow after induction regimen and at week 16 of starting treatment with a sensitivity equal to or greater than 10⁻⁴ by multiparameter flow cytometry, and relapse was defined as BM blasts greater than 5% after a CR or an extramedullary relapse [17]. Overall survival (OS) was defined as the length of time measured in months from diagnosis to death resulting from any cause. AKI was defined and graded in accordance with the most recent published guidelines [18].

Treatment regimens at our institution were fitted to previously published protocols: hyper-CVAD [19], a modified CALGB 10403 [20], and an institutional protocol (HOP 0195/0612) [21]. Rituximab was added if the percentage of CD20-positive blasts was equal to or exceeded 20%. All patients received prophylaxis with allopurinol (as rasburicase is not available in the country) with dosing according to renal function plus intravenous hydration.

Baseline characteristics were summarized as medians with interquartile ranges and percentages for descriptive purposes. Differences between categorical variables were assessed using the χ² test or Fisher’s exact test. Differences between medians were analyzed using the Mann-Whitney U test. Risk factors were assessed by calculating odds ratio (OR) along with 95% confidence interval (CI). A multivariate analysis with a logistic regression model was performed including the variables that were clinically and/or statistically significant. Time-to-event variables were estimated using the Kaplan-Meier method, and comparisons among the prognostic subgroups were analyzed using the log-rank test. Statistical significance was determined as p < 0.05. All analyses were performed using the SPSS software, Version 22 (SPSS Inc., Chicago, IL, USA).

We included 138 patients for the analysis. Seventy-three patients (52.9%) were male, and the median age at diagnosis was 34 years (range, 18–72 years). One hundred and thirty-one patients had a diagnosis of B-ALL (94.9%). Twenty-four patients (17.4%) had a medical history of diabetes, 17 patients (12.3%) of hypertension, and 2 patients (1.4%) of cirrhosis. Thirty-two patients (23.2%) had high-risk cytogenetics, of which 18 were philadelphia chromosome-positive ALL (Ph+ ALL) and 3 patients had MLL rearrangements. The remaining baseline characteristics, including blood and serum values at diagnosis, are shown in Table 1. The most common treatment regimen was hyper-CVAD (54 patients, 39.1%), followed by a modified CALGB 10403 (52 patients, 37.7%). Rituximab was added to induction therapy in 43 patients (31.2%).

The median follow-up of the cohort was 54 months (range, 5–98 months). Out of the 138 patients, 115 (83.3%) achieved CR after their first induction regimen. Among them, 55 patients experienced at least one relapse, accounting for 39.9% of the cohort or 47.8% of those who achieved CR.

During the induction treatment regimen, 97 patients (70.3%) developed febrile neutropenia, 48 patients (34.8%) developed septic shock, and 39 patients (28.3%) were transferred to an intensive care unit (ICU). In patients who survived induction therapy, the median length of hospital stay during induction was 29 days (range, 2–91 days).

TLS occurred in 42 patients (30.4%), of whom 21 patients (15.2%) had L-TLS and 21 patients (15.2%) had C-TLS. Hyperphosphatemia was the most common laboratory abnormality present in all patients with TLS, followed by hyperuricemia in 32 patients (76.1%) with TLS. In the C-TLS group, all patients developed AKI, of which only one required renal replacement therapy (RRT); 2 patients developed some type of cardiac arrhythmia, of which only one developed cardiogenic shock; and 2 patients developed some type of seizure. TLS findings are shown in Table 2.

In a univariate analysis, the occurrence of TLS was statistically associated with a history of hypertension (OR, 3.97), age 50 years or greater (OR, 2.42), an LDH 3 times greater its ULN value (OR, 3.76), and a WBC count equal to or greater than 50 × 10⁹/L (OR, 3.17). In a multivariate analysis, both WBC count greater than 50 × 10⁹/L (OR, 2.95) and LDH 3 times greater its ULN value (OR, 3.41) remained statistically associated with TLS occurrence. CI and p values are shown in Table 3. Forty-four patients had a WBC count greater than 50 × 10⁹/L, of which 21 (47.7%) developed TLS, and 41 patients had an LDH 3 times greater than its ULN value, of which 21 (51.2%) developed TLS. Presenting both variables conferred a sensitivity of 39.4%, a specificity of 92.5%, a positive predictive value of 71.4%, and a negative predictive value of 76.5% to develop TLS.

C-TLS development was significantly associated with being male (OR, 3.36), age 50 years or greater (OR, 3.86), hyperuricemia equal to or greater than 7.5 mg/dL at diagnosis (OR, 16.42), LDH 3 times greater than its ULN value (OR, 5.25), and WBC count equal to or greater than 50 × 10⁹/L (OR, 2.8). In a multivariate analysis, hyperuricemia (OR, 13.55), being male (OR, 6.46), and LDH 3 times greater than its ULN value (OR, 5.4) remained statistically associated with C-TLS development. CI and p values are shown in Table 4. Presenting these three variables together conferred a sensitivity of 47.3%, a specificity of 95.5%, a positive predictive value of 69.2%, and a negative predictive value of 89.4% to develop C-TLS.

At last follow-up, 80 patients (58%) had died, of which 25 deaths (18.1% of the cohort, 31.2% of all deaths) occurred during induction treatment. The median OS of the entire cohort was 20 months (95% CI, 14.88–25.11) with a 3-year OS of 36.6%. The median OS of patients who developed C-TLS was 12 months (95% CI, 3.71–20.28) with a 3-year OS of 19.4%, which was lower than that of patients who developed only L-TLS with a median OS of 19 months (95% CI, 0–44.76) and a 3-year OS of 30.7%, and lower than patients who did not develop TLS with a median OS of 23 months (95% CI, 13.78–32.21) with a 3-year OS of 41% (Log Rank test χ² = 8.28, p = 0.016). The Kaplan-Meier survival curves are shown in Figure 1.

TLS and C-TLS were associated with a higher risk of mortality for any cause (HR 1.61, 95% CI: [1.03–2.53], p = 0.035) and HR 2.12, 95% CI: [1.22–3.7], p = 0.007), respectively. However, this association did not remain statistically significant in a multivariate analysis that accounted for high-risk disease features, treatment with a pediatric-inspired regimen, and measurable residual disease status after induction regimen and after week 16 (univariate and multivariate Cox regression analyses for mortality are shown in Tables 5 and 6, respectively).

C-TLS was associated with a higher likelihood of death during induction therapy (HR 3.63, 95% CI: [1.6–8.23], p = 0.002), and this association remained statistically significant in a multivariate analysis adjusted for factors including age greater than 55 years, febrile neutropenia, and treatment received (HR 2.65, 95% CI: [1.13–6.21], p = 0.024). Also, TLS was associated with a higher chance of admission to an ICU during induction therapy (HR 1.95, 95% CI: [1.04–3.69], p = 0.038), although the association was not significant in a multivariate analysis adjusted by statistically significant factors (BMI ≥30, a history of hypertension, WBC count ≥100 × 10⁹/L, and febrile neutropenia). Univariate and multivariate Cox regression analyses for mortality during induction therapy and admission to an ICU during induction therapy are shown in Tables 7 and 8, respectively.

A total of 31 patients (22.5%) presented with AKI during induction therapy, of which 21 fulfilled criteria for TLS. Most of them had grade 1 (80.6%), while 6.4 and 12.9% had grades 2 and 3, respectively.

Patients in whom AKI was documented during induction had a shorter median OS (12 months, 95% CI: [4.97–19.02]) compared with patients who did not develop AKI (28 months, 95% CI: [17.98–38.01]), with a 3-year OS of 20.1 and 41.4%, respectively (Log Rank test χ² = 11, p = 0.001). The Kaplan-Meier survival curves are shown in Figure 2. Also, AKI during induction therapy was associated with a higher chance of mortality for any cause (HR 2.18, 95% CI: [1.35–3.53], p = 0.003), a higher chance of admission to an ICU during induction therapy (HR 2.43, 95% CI: [1.24–4.75], p = 0.009), and a higher chance of mortality during induction (HR 5.35, 95% CI: [2.42–11.83], p < 0.001), although only the last two remained statistically significant after adjusting for statistically significant factors (univariate and multivariate Cox regression analyses with HR, CI, and p values for AKI are shown in Table 9).

This retrospective study, conducted at a tertiary care center in Mexico City, investigated TLS in a cohort of 138 adult patients with a new diagnosis of ALL in their first induction treatment regimen over a period of seven years. We found that the prevalence of TLS in this population was 30.4%, with half classified as C-TLS and the other half as L-TLS. AKI was found in all cases of C-TLS, whereas cardiac arrhythmia or seizure occurred only in 2 patients each, always accompanied by AKI. Variables associated with the development of TLS in a multivariate analysis were a WBC count greater than 50 × 10⁹/L and LDH level 3 times greater its ULN value, and with the development of C-TLS were hyperuricemia at admission, being male, and LDH 3 times greater its ULN value. Patients who developed TLS and C-TLS had a shorter median OS than patients who did not develop these complications. Both TLS and C-TLS were associated with mortality for any cause in a univariate analysis, TLS was also associated with a higher chance of admission to an ICU during induction therapy in a univariate analysis, and C-TLS was associated with a higher probability of death during induction, and this association remained statistically significant in a multivariate analysis. Another exploratory finding was the occurrence of AKI during induction therapy, which was documented in 22.5% of patients, and this finding was also associated with worse outcomes.

To the best of our knowledge, this is the first study to explore TLS, factors associated with its development, and its relationship with the outcomes in adults with ALL, as previous research has predominantly focused on the pediatric population [22, 23]. The largest pediatric analysis found that TLS occurred in 79 of 5,537 children in a Chinese registry [23], a prevalence much lower than what we reported, which could be explained by the higher prevalence in the adult population of comorbidities associated with the loss of renal function such as hypertension and overweight, and by age itself [24‒26].

We identified different thresholds for the LDH value and WBC count compared to those proposed by the expert panel [4], which may better identify patients at higher risk of developing TLS. Additionally, hyperuricemia at diagnosis had the strongest association with the occurrence of C-TLS, and this relationship had been explored before in adults with acute myeloid leukemia [12‒14]. In our population, 71.4% of the patients with a WBC count greater than 50 × 10⁹/L and an LDH 3 times greater than its ULN value had TLS diagnosed, and 69.2% of the male patients with an LDH 3 times greater than its ULN value and hyperuricemia had C-TLS diagnosed. In contrast, of the 69 patients with a high risk for TLS development according to Cairo-Bishop stratification, only 43.4% had TLS diagnosis, showing the low precision of the current risk stratification tool in our population and the unmet need to better identify patients at risk.

According to our results, the importance of identifying patients at a higher risk of developing TLS and C-TLS lies on the finding that a shorter median OS was observed in this group, and the occurrence of the latter was also associated with a higher risk of death during the induction regimen in a multivariate analysis, probably due to AKI during induction since all patients with C-TLS had AKI in our study population. Retrospective analyses [9, 27, 28] of patients with hematological malignancies have reported higher in-hospital mortality, 6-month mortality, and lower CR rate in patients who develop TLS and AKI. Also, a large analysis of 22,875 TLS hospitalizations (more than half were due to hematological malignancies) from a large US nationwide reported that RRT was needed in 12% of them, and it was associated with greater odds of mortality and adverse discharge [10]. A prospective study also reported that developing AKI within 3 weeks of hematological malignancy diagnosis (12% were ALL) was associated with a decline in kidney function within 1 year [29]. All this evidence suggests that patients who present with kidney injury in the context of TLS (or in any scenario) are in a higher risk of adverse outcomes. RRT was only needed in 9.5% of patients with C-TLS in our population, and over 80% of AKI were grade 1, proposing that even mild injury to kidney function detected by slight changes in serum creatinine may confer inferior outcomes, probably driven by the loss of renal reserve function, more toxicity in further chemotherapy courses, and chemotherapy dose-adjustment due to kidney injury leading to lower drug dose administration and subsequent chemotherapy cycles’ delay.

Although out of the scope of our work, an important finding was that treatment with a pediatric-inspired regimen was associated with a lower risk for mortality for any cause and during induction, remaining statistically significant in a multivariate analysis, consistent with evidence published during the last decade when treating adolescents and young adults [30]. Induction-related mortality in limited resource settings is a determinant of OS. In our cohort, 18.1% of patients died during induction therapy, which is similar to other reports from other lower and median income settings [31‒33], so any strategy that might reduce this outcome, especially in our environment, is a need. Given that AKI and C-TLS were associated with induction-related mortality, interventions aimed at reducing these complications could potentially reduce induction-related mortality and, consequently, overall mortality. Besides what has been investigated about rasburicase [34, 35] (which is not available in our country), there has not been any recent research to try standardizing hydration, diuresis protocols, and potentially new therapeutic approaches in this group of patients. Phosphate binders have been used in case reports in children with TLS [36, 37], although this finding should be tested in prospective trials with a larger population. Although there is no published evidence favoring their use, the use of phosphate and potassium binders are part of some institutions protocols for preventing and treating TLS, so any information on the impact of these interventions is needed. With newer technologies to guide hydration like point-of-care ultrasound are now commonly being used, it should be of interest to see if these kinds of approaches could lower the incidence of AKI during induction in patients with hematological malignancies.

The retrospective nature of our investigation and that it was conducted exclusively at one center limit the ability to generalize our results. Attempts must be made to prospectively follow a cohort of newly diagnosed ALL patients who develop TLS and/or AKI during induction therapy and determine the real impact of our findings and population at risk.

TLS occurred in almost one-third of patients from a cohort of newly diagnosed ALL patients from Mexico City, and half fulfilled the criteria for C-TLS. We report different thresholds for the LDH value and WBC count that may put patients at risk to developing these complications, as well as other previously described factors associated with TLS in other hematological malignancies such as hyperuricemia. The median OS of patients who developed TLS and C-TLS is shorter than that of those who did not, and C-TLS is independently associated with a greater risk of mortality during induction therapy. AKI during induction also confers a shorter median OS and is associated with worse outcomes as previously described in other hematological malignancies. Efforts should be made to design prospective studies that better define the patient populations at risk of developing these complications, assess their impact, and systematize interventions while exploring potential new approaches.

Ethical approval is not required for this study in accordance with local or national guidelines. Patient consent was not required in accordance with local or national guidelines. It meets all ethical requirements.

Roberta Demichelis-Gómez receives honoraria from AbbVie, AMGEN, Astellas, Gilead, Pfizer, and TEVA as a speaker and/or part of the advisory board.

This research did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.

F.R.O. and R.D.G. conceived and designed the study, performed the statistical analysis, and analyzed and interpreted the data; F.R.O., F.G.L., and A.M.C. provided the study material and/or patients; F.R.O. drafted the manuscript; and all authors reviewed and approved the manuscript.

All data generated or analyzed during this study are included in this article and the supplementary material. Further inquiries can be directed to the corresponding author (R.D.G.). All the authors had access to the clinical data. Individual participant data will not be shared.