An adverse drug reaction (ADR) may be defined as "an appreciably harmful or unpleasant reaction resulting from an intervention related to the use of a medicinal product; adverse effects usually predict hazard from future administration and warrant prevention, or specific treatment, or alteration of the dosage regimen, or withdrawal of the product".¹ An ADR is an important cause of illnesses and death in pediatric cancer patients and cancer survivors.² It has been estimated that almost 75% of cancer survivor patients have suffered at least one ADR during their treatment, nearly 40% of them have had at least one life-threating reaction and 3.2% of these patients have died due to one ADR.³

Children are more susceptible to develop an ADR due to major differences in pharmacokinetics and pharmacodynamics of drugs used in pediatrics, off-label use, and because their illness unlike in adults. Additional risk factors such as polypharmacy, prolonged hospital stays, and critical conditions increase the likelihood of developing an ADR.⁴

Anthracyclines are drugs commonly used in pediatric patients with cancer and are a part of several chemotherapy regimens. These drugs are listed among the World Health Organization’s model list of essential medicines and include epirubicin, doxorubicin, daunorubicin, and idarubicin.⁵ Despite the efficacy of anthracyclines, their use is limited because an ADR is likely to occur, the most important cardiomyopathy. In fact, anthracyclines-associated reactions require special care and drug treatment. The anthracycline-associated acute toxicity includes myelosuppression, mucositis, nausea and vomiting, diarrhea, and alopecia. The incidence of these reactions has been documented as case reports. These reactions, if present, might restrict the use of these drugs even in standard doses. The development of anthracycline-associated cardiomyopathy is the most critical long-term ADR and it is the major reason for limiting anthracycline use.^6,7 Nevertheless, it is not clear how the risk factors or the doses are related to damages.⁸

The anthracyclines treatment, used as many other chemotherapy agents, is not specific to cancer cells; thus, it has the potential to damage proliferating cells in healthy tissues. Nevertheless, it is common for health professionals to accept all ADRs as normal (or expected) events. Therefore, it is necessary to use pharmacovigilance to know the incidence, pattern, causality, and assessment strategies of prevention.

Pharmacovigilance is defined as the science and associated activities related to the detection, assessment, understanding, and prevention of adverse effects or any other drug-related problem.⁹ Pharmacovigilance actions (detection, notification, and registry of an ADR) guarantee the drugs are safely used. According to the International Conference of Harmonization, it is recommended to apply local and international guidelines in pharmacovigilance to fulfill these actions.¹⁰

Therefore, the aim of this retrospective cohort study was evaluating the incidence, severity, and causality of ADRs associated with anthracycline-based chemotherapy in Mexican cancer pediatric patients, based on pharmacovigilance guidelines.

A retrospective cohort study was conducted to integrate pediatric patients who received anthracyclines at two major pediatric institutes: Hospital Infantil de México "Federico Gómez" (HIM) and Hospital Pediátrico (HP) at Centro Médico Nacional from Instituto Mexicano del Seguro Social. This was a four-year study, which began when the treatment started. The sample consisted of 99 patients who met the following inclusion criteria: children diagnosed with cancer, both genders, children treated with anthracyclines, and the affected children who had finished exposure to anthracyclines based-chemotherapy. The exclusion criteria were: patients who continue with cancer treatment or have another unknown source of anthracycline administration, patients who have another type of malignancy, and patients who present a disease that might be confused with an ADR. The definitions of ADRs used in this study are based on the Common Terminology Criteria for Adverse Events. For all, we used the last version, and in the case of left ventricular systolic dysfunction and vomit were also classified with the previous version in order to obtain the specific ranges of each severity grade.¹¹

We identified the oncology patients through the Division of Statistics and Informatics of each hospital. All data were obtained from medical records and these were reviewed for each case and intentionally the information regarding ADRs associated with anthracycline-based chemotherapy. The causality assessment of the ADRs was evaluated using the Naranjo algorithm.¹² This latter is one of the most used to assess the causality of an ADR. This algorithm gives us the probability that an event is associated with a drug. That score allows us to classify an ADR as doubtful (0), possible (1-4), probable (5-8), and definitive (>9). The severity of these reactions was determined according to CTCAE.¹¹ The data from medical records were collected in a specific profile containing the patient’s background and information of the ADRs. Incidence of the ADR anthracycline-based chemotherapy was obtained through the relation between the number of patients who developed each of the reactions, and the total patients in the cohort study.

Statistical Analysis

The sample size was calculated using the one proportion formula, with α of 99% and 8% of precision to find an incidence of left ventricular systolic dysfunction of 10%.¹³ Under these conditions, we calculated a sample size of 94 patients.

Quantitative variables with no normal distribution such as age are presented as median and interquartile (25-75) range.

Categorical variables such as gender, diagnosis, radiotherapy, nutritional status, and information about the treatment are presented as number (n) and percent (%). The cumulative doses of the different anthracycline were matched by calculating the equivalence to the cumulative dose of doxorubicin using the conversion factors described by Barnard.¹⁴

Mann-Whitney U and χ² or Fisher’s exact test were utilized to estimate the differences between hospitals (HP vs. HIM). IBM SPSS version 22.0 (Armonk, NY, IBM Corp) was used for statistical analysis.

Ethics

This protocol was approved by the Research, Ethics and Biosafety Committees in both institutions HIM and PH IMSS (registration code: HIM-2013-062 SSA 1091 IMSS: R-2013-785-040). All potential patients were invited to the study and asked to sign an informed consent form.

Demographic characteristics of the sample

A total of 99 patients were included in this study; Table 1 shows the demographic characteristics. The median age was 5.17 years (interquartile, 2.8-9.3). A little over half of the patients were male (55.5%). The most frequent diagnoses were leukemia and lymphoma with 61 (61.6%), using daunorubicin (37.4%), epirubicin (31.3%) and doxorubicin (27.3%). Only 28 (28.3%) patients received radiotherapy. Malnutrition status was found in 18 (18.2%) patients. There were two variables with statistical significance between hospitals: cumulative doses of anthracycline (>450 mg/m²) with a value of p=0.007 and the use of cisplatin as concomitant drug (p=0.002).

Adverse reactions associated with anthracyclines-based chemotherapy

Ninety-two percent out of the 99 patients included in the study had at least one ADR associated with chemotherapy. A total of 718 ADRs were found, 354 (49.3%) of them were associated with anthracyclines-based chemotherapy as either probable and definite, according to the Naranjo Algorithm. The incidence of ADRs related to anthracyclines-based chemotherapy were: febrile neutropenia 67.7%, thrombocytopenia 39.4%, mucositis 37.4%, anemia 33.3%, vomit 15.2%, left ventricular systolic dysfunction 12.1%, and diarrhea, 7.1% (Table 2 and Graphic 1). There was a difference between hospitals in the incidence of febrile neutropenia (p=0.005), left ventricular systolic dysfunction (p=0.01), and diarrhea (p=0.038).

Three organ systems were adversely affected by anthracyclines-based chemotherapy: gastrointestinal, blood and lymphatic, and cardiac systems. The majority were severe (Grade 3) according to CTCAE. One of the ADRs was highly associated with death (see Table 3).

The number of ADRs associated with anthracycline-based chemotherapy per patient was a median of 4 ADRs/patient, and the range was from 1 to 28 ADRs/patient. The majority had only one ADR, but some patients had more than 4 ADRs (Graphic 2).

Anthracyclines are among the most potent chemotherapeutic agents; unfortunately, like other cancer drugs, they are not specific for cancer cells, and their toxicity is associated with the impact of these drugs with other organs and systems. The period of presentation is generally within one to two weeks for hematologic reactions, but years for cardiovascular reactions. More than 50% of children suffering from cancer are treated with these drugs.¹⁵ The ADRs associated with anthracycline-based chemotherapy can compromise the success of therapy, by the association with complications that require hospitalization or may result in dead. ADRs are an important issue for health systems since they entail an increase of days of hospitalization and emergency service visits or admissions. Pediatric patients are far more susceptible to serious ADRs because they have smaller body sizes and, because of that, smaller body surface than adults, and especially because they have many differences in the way they absorb, distribute, metabolize, and excrete drugs.^16,17

In this study, we investigated the incidence and severity of ADRs in two different tertiary care hospitals in Mexico. We were able to show that 92% of the cases had at least one ADR associated with chemotherapy, and a total of 718 ADRs were found, 354 (49.3%) of them were associated with anthracyclines-based chemotherapy. ADRs are a critical clinical problem in cancer pediatric patients as reported by Elzagallaai AA et al. because an increase of ADRs (from 5% to 50%) with chemotherapeutic agents was found.¹⁸ It is noteworthy that not all ADRs are reported at hospitals, and this might be a factor in variation of ADR incidence around the world. Morales Ríos O et al. found that 97% of ADRs are correctly identified, but only 6.1% are reported. However, after a pharmacist’s intervention that included group informative talks on pharmacovigilance, reminders, and feedback to the personnel, reports of ADRs increased to 41.7%, showing that the intervention had a positive effect on reducing under-reporting.¹⁹

Similarly to this study, Parande F. et al. observed 332 adverse events in 55 cancer pediatric patients, and most of them were suspected to have been caused by the administration of daunorubicin and methotrexate²,suggesting that these ADRs should be monitored through pharmacovigilance activities to avoid compromising the patient’s life.

Rinke M. et al. found that 48% of pediatric chemotherapy error reports originated in the phase of drug administration; incorrect dose/quantity (22.9%) was the most common error.²⁰ Another recurrent mistake is confusing the values of body surface area with body mass index, or to copy the values from the previous cycle of chemotherapy. These errors can be prevented by drug-dose checking and guide dose algorithms as shown by Gandhi et al.²¹ and by administering and checking the value of cumulative doses until finishing the treatment. In our study, all doses of each anthracyline were matched calculating the equivalence to doxorubicin doses, and some patients had cumulative doses of anthracycline greater than 450 mg/m².

In general, the incidence of ADRs associated with anthracycline-based chemotherapy was high in both hospitals; the only difference was related to the incidence of febrile neutropenia, left ventricular systolic dysfunction, and diarrhea. Neutropenia is a common, dose-dependent toxicity, usually occurring in a reversible manner, with low mortality rates.²² The nadir occurs 10-14 days after administration.²³ The incidence found in this study of febril neutropenia (67.7%) associated with anthracycline is higher compared with adults and even with other studies in pediatric patients. Minuk LA. et al. found that the incidence of febrile neutropenia was 0.5% in pediatric patients treated with doxorubicin, bleomycin, vinblastine, and dacarbazine.²⁴ Castelán-Martínez O. et al. found that anthracyclines were a risk factor (OR=19.44 [95% CI, 5.40-70.02]) for the presence of febrile neutropenia in pediatric patients treated with cisplatin-based chemotherapy.²⁵ In this study, all the patients received colony-stimulating factors at least once, as well as antibiotic prophylaxis. Nevertheless, the incidence of febrile neutropenia should be considered to evaluate the efficacy of prophylaxis treatment.

Thrombocytopenia and anemia are two severe ADRs that directly impact the quality of life in cancer patients and almost always call for treatment and hospitalization. Transfusion is the most effective treatment in these cases. Approximately one million units of blood are transfused in cancer patients each year in the United States.²⁶ However, the use of transfusions carries the risk of other reactions such as transfusion-associated sepsis. Mucositis is a reaction that can suddenly impair the state of child’s health; the patient stops eating because of pain and discomfort caused by mouth ulcerations. In the hospitals where the study was performed, the treatment for mucositis is based on mouthwashes with sodium bicarbonate, parenteral nutrition, antibiotic and antifungal prophylaxis, and constant monitoring of the total neutrophil count. This reaction is one of the most expensive and also has a high probability of turning into major complications such as anal and gastrointestinal mucositis.²⁷

Vomiting was less frequent than in other studies. An incidence of 84% has been reported²⁸,suggesting that there is an under-reporting of this reaction or a good effectiveness of prophylactic treatment. Le Baron S. et al. found that for anthracyclines-based chemotherapy there was more severe emesis for courses with antiemetics than for those with none.²⁹ So, it is relevant considering the adequate prophylaxis for this reaction.

Most of the authors only report cardiomyopathy as documented and relevant reactions associated with anthracyclines-based chemotherapy. In this study, we found an incidence of left ventricular systolic dysfunction of 12.1% and the majority was mild or grade 1, according to CTCAE. The incidence in this study was less than the one reported by Lipshultz SE et al., who found an incidence of 47% in a cohort of cancer survivors.³⁰ Sorensen K. et al. showed that 23% of acute lymphoblastic leukemia survivors had abnormalities determined by echocardiography. Increased afterload caused by reduced left ventricular posterior wall thickness occurred in 25 (21%) patients, and reduced contractility occurred in 2 patients (2%), suggesting that it is necessary to follow this reaction even years after treatment.³¹ It has been documented that this test could be made with another test like troponins values or nuclear magnetic resonance to ensure that patients do not have a non-clinical myocardial dysfunction.³²

Another factor to consider is the latency of this ADR. It has been documented that there have been cases of cardiotoxicity 20 years after anthracycline exposure.³³ In our hospitals, when a child became an adult, he or she is transferred to an adult hospital, which makes the follow-up difficult; in many cases, the patient is lost for follow-up. ADRs associated with chemotherapeutic drugs decrease the quality of life and increase mortality as well as the health care budget. The ADR profile is less reported, and the situation is even worse in Mexico. The variability in the temporality, severity, and frequency of the reactions associated with anthracyclines-based chemotherapy makes it vital that we monitor, analyze, and report on them to improve effective actions to prevent them.

We were able to show that besides left ventricular systolic dysfunction —as the most relevant related to anthracyclines-based chemotherapy ADR—, other frequent ADRs were found in this study. These results must be considered to update the safety profile of anthracyclines-based chemotherapy in pediatric population. All the ADRs are important because of their high incidence and severity in pediatric populations.

Nails discoloration, alopecia, and hyperpigmentation of the skin associated with anthracyclines-based chemotherapy were not found in this study, showing that most of these ADRs are assumed to be expected and are not documented.

In terms of causality, we used the Naranjo algorithm in order to analyze the probability that the treatment may be associated with the event. There is not a gold algorithm to use, but this one is the most used in pharmacovigilance. As we obtained values of probability between probable and definite, it was difficult to attribute the events only to anthracyclines-based chemotherapy. One limitation of using this algorithm is that it does not explain the causality with more concomitant treatments.

Following the guidelines of pharmacovigilance enable detecting, analyzing, and preventing ADRs. Consequently, it is necessary that pharmacovigilance activities be incorporated in hospital systems, not like a procedure but as a part of everyday activities of professional health. In Mexico, there have been substantial advances in pharmacovigilance regulation, but this study shows that numerous ADRs are not reported and they are taken as normal events. Like all retrospective studies, we are limited in that the information depends directly on the clinical record. Further, we do not control the period when tests to show the ADRs are taken