Chemotherapy-induced toxicities and associated factors among patients with gastrointestinal tract cancer at two tertiary care hospitals in Ethiopia: a prospective observational study

Abstract

Background

Gastrointestinal tract (GIT) cancers are a leading cause of cancer-related death worldwide. Chemotherapy, a cornerstone of treatment, often causes toxicities that can compromise clinical outcomes. However, evidence on the prevalence and risk factors of chemotherapy-induced toxicity among GIT cancer patients in Ethiopia remains limited. Therefore, this study aimed to assess the prevalence and associated factors of chemotherapy-induced toxicity among GIT cancer patients at two Tertiary Care Hospitals in Addis Ababa, Ethiopia.

Methods

A hospital-based prospective observational study was conducted among patients enrolled prior to chemotherapy initiation and followed biweekly until treatment completion. Adverse drug reactions (ADRs) were collected from patient interviews and medical records, and assessed using the Naranjo algorithm, Hartwig-Siegel scale, and National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) version 5. Data were analyzed using descriptive statistics and logistic regression in SPSS version 26, with P < 0.05 considered statistically significant.

Results

The mean (± SD) age of patients was 51.59 (± 11.18) years. A total of 1499 ADRs were reported, of which 47.4% (n = 710) were hematological toxicities. Based on the Naranjo scale, 54.8% of the ADRs were classified as probable, while the Hartwig—Siegel scale severity score classified 53.7% as mild ADRs. According to NCI-CTCAE v5.0 grading score, 37.1% of the ADRs were graded ≥ 3. Poor pathological grade (AOR: 4.410, 95% CI: 1.823–10.666, p = 0.001), low baseline hemoglobin (AOR: 4.456, 95% CI: 2.021–9.824, p = 0.001) and low absolute neutrophil counts (AOR: 2.237, 95% CI: 1.061–4.717, p = 0.034) were significantly associated with severe ADRs. Granulocyte colony-stimulating factor (G-CSF) use was independently associated with lower odds of high-grade ADRs (AOR = 0.508, 95% CI: 0.260–0.995, p = 0.048).

Conclusion

Chemotherapy-induced toxicities were highly prevalent among GIT cancer patients, with a substantial proportion experiencing severe toxicities. Poor pathological grade, low baseline hemoglobin, and low absolute neutrophil count were significantly associated with high-grade toxicities, while G-CSF use was protective. These findings highlight the need for improved toxicity management and personalized treatment strategies in Ethiopian oncology practice.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12885-026-16554-6.

Backgrounds

Cancer is the second leading cause of death globally, causing an estimated 9.6 million deaths in 2018, or about one in six deaths [1, 2]. Over the past two decades, cancer incidence has risen worldwide, with nearly 70% of cancer‑related deaths now occurring in low‑ and middle‑income countries, where the burden is expected to increase [3–5]. Globally, lung cancer is the leading cause of cancer death, followed by colorectal, liver, stomach, and breast cancers, with gastrointestinal tract (GIT) cancers accounting for 36.2% of cancer deaths [6, 7]. In Africa, colorectal cancer ranks fifth, with higher prevalence in northern Africa (8.8%) than sub‑Saharan Africa (5.9%) [8, 9]. In Ethiopia, 2019 saw approximately 53,560 new cancer cases, with gastrointestinal cancers, particularly colorectal cancer, rising at incidences of 8.5 and 6.3 per 100,000 in men and women, respectively [1, 10].

The treatment of gastrointestinal tract cancer consists of surgery, chemotherapy, radiation, and biological modulators. The most common chemotherapy regimens include 5-FU-based combinations, such as FOLFOX, FOLFIRI, FOLFOXIRI, CAPOX, and XELIRI [8, 9].Despite the growing use of chemotherapy in sub‑Saharan Africa, data on chemotherapy‑induced toxicity and its impact on treatment delivery and quality of life remain limited, particularly in resource‑limited settings [11]. Chemotherapy‑induced toxicities such as neutropenia, febrile neutropenia, nausea and vomiting, and gastrointestinal symptoms are major causes of treatment interruption and reduced quality of life, yet comprehensive toxicity profiles from sub‑Saharan African cohorts are scarce [12]. In low‑ and middle‑income countries, including Ethiopia, chemotherapy‑related toxicities are often under‑documented, limiting optimization of supportive care and dose intensity in patients with GIT cancers [13]. Adverse drug reactions (ADRs) from cancer treatment cause significant morbidity, strongly impact the quality of life, and frequently impair the delivery of optimal therapy, with common toxicities including alopecia, nausea and vomiting, myelosuppression, cardiac toxicity, hemorrhagic cystitis, mucositis, hot flashes, electrolyte imbalance, neuropathy, and deep vein thrombosis [14]. The majority of patients who are receiving chemotherapy for GIT cancer are affected by treatment-related ADRs, and the most commonly reported symptoms are fatigue (88%), diarrhea (75%), constipation (73%), and vomiting (58%) [3, 8, 15, 16]. Chemotherapy-induced nausea and vomiting (CINV) is associated with a high risk (> 90%) with IV cisplatin, whereas there is a low risk (< 10%) with IV rituximab [16].Chemotherapy-induced hematologic toxicities can be life-threatening, with neutropenia posing the highest risk due to the potential for severe infections [17]. For regimens with a predicted risk of febrile neutropenia > 20%, primary prophylaxis with granulocyte colony-stimulating factor (G-CSF) is recommended. However, prior to this study, data on the incidence of chemotherapy-induced neutropenia, febrile neutropenia, and G-CSF use in sub-Saharan Africa, including Ethiopia, were limited. Similarly, comprehensive data on non-hematologic chemotherapy-related toxicities in the region were scarce, limiting evidence-based optimization of supportive care [18–20].

This study provided comprehensive data on chemotherapy-induced toxicities and associated factors among Ethiopian patients with GIT cancer. The findings identified high-risk patients and specific risk factors, enabling targeted interventions to improve patient outcomes. They also provide evidence to guide resource allocation for effective toxicity management in Ethiopia, reducing morbidity and healthcare costs while enhancing patients’ experiences and quality of life.

Methods

Study area and period

This study was conducted at the adult oncology inpatient units of Tikur Anbesa Specialized Hospital (TASH) and St. Paul’s Millennium Medical College (SPHMMC) Addis Ababa, Ethiopia. TASH, established in 1972 in Addis Ababa, is the country’s primary referral center, treating over 6,000 new cancer patients annually and currently serving more than 10,000 patients. It began training residents in clinical oncology in 2013, with 36 resident enrolled. SPHMMC founded in 1968 and formed as a medical college in 2007, established oncology unit on August 1, 2018, as the second hospital in the country to offer cancer treatment. It currently employs two oncologist, 12 general practitioners, 14 nurses, and serves around 800 cancer patients. The study was conducted from June 20, 2023 to February 15, 2024.

Study design and population

A hospital-based prospective observational study was conducted. Patients were enrolled prior to the initiation of chemotherapy and followed throughout their treatment until completion. During each cycle, participants were assessed every two weeks, and ADRs were documented using structured data collection tools.

Inclusion and exclusion criteria

Patients aged 14 years or older with histologically or clinically confirmed GIT cancer who initiated chemotherapy during the study period were included. All participants were admitted to the adult oncology units of TASH or SPHMMC and provided written informed consent. Patients were excluded if they had incomplete medical records, missing baseline clinical or laboratory data, or inadequate documentation of chemotherapy administration. Those who declined participation, received chemotherapy outside the study hospitals, had severe comorbidities, pre-existing non-cancer hematologic disorders, or were unable to attend scheduled follow-ups were also excluded.

Sample size determination

All adult patients with GIT cancer attending TASH and SPHMMC during the study period were recruited and followed until treatment completion. Most patients completed first-line chemotherapy in six cycles, lasting approximately 6–8 months. Participants were enrolled using non-probability consecutive sampling. A total of 209 adult GIT cancer patients receiving chemotherapy were assessed for eligibility, and five patients from TASH were excluded. The final analysis included 204 patients (TASH: 148; SPHMMC: 56) (Fig. 1).

Fig. 1.

CONSORT-Style flow diagram of study participants

Data collection and management

A structured data abstraction form was developed based on a review of the literature, capturing socio-demographic, clinical, pathological, and treatment-related variables. Data collectors received training on the tool and standardized collection procedures. The tool was pretested on 10 patients not included in the study, and necessary revisions were made prior to data collection.

Patient-reported data, including socio-demographic characteristics and non-hematologic adverse drug reactions (ADRs) such as nausea, vomiting, diarrhea, constipation, dysphagia, and alopecia, were collected prospectively during follow-up visits using a structured data collection form. Clinical, pathological, laboratory, and treatment-related variables—including cancer diagnosis, tumor grade, chemotherapy regimen, baseline hemoglobin level, absolute neutrophil count (ANC), and granulocyte colony-stimulating factor (G-CSF) use—were extracted from patients’ medical records by trained data collectors.The data collection process followed a systematic multi-source approach under the close oversight of the principal investigator. When discrepancies identified by the data collectors could not be resolved, additional clarification was obtained through consultation with oncology nurses, treating physicians, or oncology residents. The principal investigator also provided continuous supervision and conducted periodic audits to ensure adherence to the study protocol and maintain data quality.

To address missing data, a complete-case analysis approach was employed. Only patients with complete cycle-level records for the key study variables were included in the final analysis, while records with missing exposure or outcome variables were excluded.

Baseline laboratory values, including hemoglobin and absolute neutrophil count, were obtained prior to initiation of chemotherapy and before the occurrence of grade ≥ 3 toxicity, ensuring temporality of exposure. G-CSF administration was recorded as prophylactic use, given before the development of severe toxicity. G-CSF administration was classified as primary prophylaxis (before chemotherapy), secondary prophylaxis (in subsequent cycles after prior neutropenia), or therapeutic (after neutropenia onset). Only primary prophylactic G-CSF was included in the regression analysis. Timing relative to neutropenia onset was verified from medical records to ensure temporality and minimize confounding by indication. A sensitivity analysis excluding patients who received G-CSF after the first episode of severe neutropenia confirmed that the protective association remained consistent.

Chemotherapy-induced toxicity assessment

Patients with gastrointestinal (GI) cancer underwent routine Clinical and laboratory evaluations, including complete blood count (CBC, prior to each chemotherapy cycle. Clinical status and treatment-related toxicities were systematically assessed by oncologists or oncology residents and documented in medical records. Trained clinical pharmacist and nurse data collectors prospectively reviewed patient charts throughout all chemotherapy cycles. NCI-CTCAE was used to grade the clinical severity of each adverse drug reaction. Hematologic and non-hematologic toxicities were graded using the NCI-CTCAE, version 5.0, with standardized application across cycles [21]. Dose intensity was calculated according to NCCN guidelines (v2.2024), and dose modifications followed institutional protocols aligned with NCCN and ESMO recommendations. At TASH and SPHMMC, hematologic toxicities prompted 20–25% dose reductions for absolute neutrophil count (ANC) < 1000/μL or platelet count < 75,000/μL, treatment delays of 1–2 weeks for ANC < 500/μL, and discontinuation after two consecutive cycles with ANC < 1000/μL. Non-hematologic toxicities of grade ≥ 3 required treatment interruption until recovery to ≤ grade 1, with permanent discontinuation for grade 4 events. Oxaliplatin-induced neuropathy warranted a 20% dose reduction for grade 2 toxicity and discontinuation for persistent grade 3 toxicity. Treatment was also discontinued in cases of cumulative delays exceeding three weeks or disease progression based on Response Evaluation Criteria in Solid Tumors (RECIST). To ensure reliability, initial toxicity grading was performed by the principal investigator, with independent review of 10–15% of cases by oncologists or oncology residents; discrepancies were resolved by consensus.

All chemotherapy-induced adverse events were systematically assessed across both hematologic and non-hematologic toxicities and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) suing standardized CTCAE grading scale (Grade 1–5). Hematologic toxicities, including anemia, neutropenia, and thrombocytopenia were identified through routine pre-cycle complete blood count (CBC) testing and graded according to CTCAE criteria. Non-hematologic toxicities were assessed comprehensively through patient interviews during follow-up visits, review of clinical documentation and physicians’ records. Rather than limiting the assessment to selected symptoms, all clinically relevant non-hematologic toxicities reported or documented during treatment were evaluated and graded according to CTCAE criteria. These included, but were not limited to, nausea, vomiting, diarrhea, constipation, dysphagia, fatigue, mucositis, peripheral neuropathy, hand-foot syndrome, and other treatment-related symptoms. To ensure consistency across patients, toxicity assessment followed a standardized data collection procedure conducted by trained data collectors under the supervision of the principal investigator. Clinical findings, laboratory results, and patient-reported symptoms were cross-checked with medical records and oncology care team documentation whenever necessary.

Causality assessment of ADRs was conducted using the Naranjo algorithm, classifying reactions as definite (≥ 9), probable (5–8), possible (1–4), or doubtful (≤ 0). The Hartwig–Siegel Severity Assessment Scale was used to classify the clinical consequences of each reaction as as mild (levels 1–2), moderate (levels 3–4), or severe (levels 5–7), based on the level of medical intervention required and the resulting patient outcomes. [22].

Study outcome variables

The primary outcome was chemotherapy-induced toxicity, evaluated at the patient level. For each participant, the highest-grade ADR experienced across all treatment cycles was recorded. Patient-level outcomes were defined as any toxicity graded ≥ 3 according to NCI-CTCAE v5.0, encompassing both hematologic and non-hematologic toxicities. This approach consolidated 1,499 cycle-level ADRs into a single binary patient-level variable (severe toxicity: yes/no) for regression analyses, ensuring a clear distinction between event-level and patient-level data.

Dependent variables included the type of chemotherapy-induced toxicity (hematologic versus non-hematologic) and its severity. Independent variables comprised socio-demographic factors (age, sex, education, residence, and marital status), clinical characteristics (cancer type, presenting symptoms, symptom duration, comorbidities, laboratory and clinical findings, cancer stage, histopathological type, and metastatic sites), and treatment-related factors (chemotherapy regimen, frequency, dose, number of cycles, treatment duration, treatment intent, and use/timing of G-CSF).

Data entry, analysis, and interpretation

Collected data were entered and analyzed using SPSS version 26. Continuous variables with normal distribution, such as age, body surface area, and body mass index, were expressed as mean ± standard deviation (SD), while non-normally distributed variables were summarized as median and interquartile range (IQR). Categorical variables, including the type and severity of hematologic and non-hematologic toxicities, were reported as frequencies and percentages.

Logistic regression analyses were performed to identify factors associated with the highest-grade chemotherapy-induced toxicity (combined hematologic and non-hematologic). Variables with p < 0.25 in bivariable analysis and clinically relevant variables were included in the multivariable model. Adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were reported, and statistical significance was set at p < 0.05. Model adequacy was evaluated using the Hosmer–Lemeshow goodness-of-fit test, and multicollinearity among independent variables was assessed to ensure the robustness and validity of the findings.

Definitions of terms

Stage of gastrointestinal cancer: diagnosis stage of cancer with pathological confirmation from stage I to stage IV [23].

ECOG Performance Status: A scale assessing the patient’s functional ability and impact of disease on daily activities, ranging from 0 (fully active) to 5 (death). Grades 1–4 indicate increasing limitation in self-care and mobility, from minor restrictions (1) to complete disability and confinement (4) [24].

Neutropenia was defined as an absolute neutrophil count (ANC) < 1500/μL, consistent with NCI CTCAE v5.0 [21]

Toxicity: Refers to the harmful effect itself, usually described in terms of type and severity, and commonly graded using the National Cancer Institute Common Terminology Criteria for Adverse Events [25].

Adverse Drug Reactions (ADRs): any response to a drug, which is noxious, unintended, and occurs at doses used in man for prophylaxis, diagnosis, or therapy.

Adverse drug event (ADE) refers to any injury caused by a medicine and it encompasses all adverse drug reactions (ADRs), (including allergic or idiosyncratic reactions) as well as medication errors (MEs) that result in harm to a patient [25]

Results

Socio-demographic characteristics of the patients

A total of 209 adult patients with GIT cancer receiving chemotherapy were recruited. Of these, 5 patients (2.5%) were excluded, leaving 204 patients for final analysis, yielding a response rate of 97.6%. The majority of patients (72.5%) were enrolled from TASH. Females comprised 51.5% of the study population. The mean ±  SD age at diagnosis was 51.59 ± 11.18 years, with 32.4% of patients aged ≥ 60 years. Most patients resided in Addis Ababa (42.6%), and 69.1% were married (Table 1).

Table 1.

Socio-demographic characteristics of GI tract cancer patients treated at the oncology unit of TASH and SPHMMC between June 2023 and February 2024, Addis Ababa, Ethiopia (n = 204)

Variable		Frequency n (%)	Mean ± SD
Study site	TASH	148 (72.5)
	SPHMMC	56 (27.5)
Gender	Male	99 (48.5)
	Female	105 (51.5)
Age in year	< 40	37 (18.1)	51.59 ± 11.18
	40–49	43 (21.1)
	50–59	58 (28.4)
	60 and Above	66 (32.4)
Marital status	Married	141 (69.1)
	Single	20 (9.8)
	Windowed	23 (11.3)
	Divorced	20 (9.8)
Educational level	No formal education	95 (46.1)
	Primary school(1–8) grade	41 (20.1)
	Secondary school(9–12) grade	31 (15.2)
	Higher education (diploma and above)	37 (18.1)
Region	Oromia	55 (27.0)
	Addis Ababa	87 (42.6)
	Amhara	30 (14.7)
	Dire Dawa	5 (2.5)
	SNNP*	25 (12.3)
	Other**	2 (1.0)
Alcohol use	Yes	37 (18.1)
	No	167 (81.9)
Cigarette smoking	Yes	35 (17.2)
	No	169 (82.8)
Khat use	Yes	40 (19.6)
	No	164 (80.4)

Abbreviations: SNNP* Southern Nations Nationalities and People, SD Standard division, TASH-Tikur Anbessa Specialized Hospital, SPHMMC-St. Paul’s Hospital Millennium Medical College; other**: Tigray (1) and Afar (1)

Clinical characteristics of the patients

Among the 204 patients, 65 (31.9%) had comorbidities primarily hypertension (10.3%,) followed by diabetes mellitus 3.4%, and retroviral infection 3.4%. Adenocarcinoma was the predominant histopathological type accounting for 72.5% of cases; of these, 72.1% had well-differentiated tumors. At baseline (cycle 1) 72.1% of patients had ECOG 0–1, which increased to 94.6% at cycle 6. The mean BSA and BMI were 1.56 ± 0.014 m² and 20.96 ± 3.6 kg/m², respectively. Stage IV disease was present in 64.7% of patients; the liver was the common metastatic site (35.6%) (Table 2).

Table 2.

Clinical characteristics of GI tract cancer patients treated at the adult oncology unit of TASH and SPHMMC between June 2023 and February 2024, Addis Ababa, Ethiopia (i = 204)

Variables			Frequency n (%)
Comorbidities		No comorbidity	139 (68.1)
		HTN	21 (10.3)
		DM	7 (3.4)
		CKD	5 (2.5)
		RVI	7 (3.4)
		Asthma	6 (2.9)
		BPH	5 (2.5)
		Other*	14 (6.8)
Duration of symptoms		≤ 6 month	88 (43.1)
		> 6 month	116 (56.9)
Histopathology cell type		Adenocarcinoma	190 (93.1)
		Squamous cell carcinoma	14 (6.9)
Pathological grading		Well differentiated	147 (72.1)
		Moderately differentiated	11 (5.4)
		Poor differentiated	46 (22.5)
Stage of cancer		Stage 3	72 (35.3)
		Stage 4	132 (64.7)
Metastasis of diseases		Yes	132 (64.7)
		No	72 (35.3)
Site of metastasis (n = 132)		Lung	18 (13.6)
		Liver	47 (35.6)
		Liver and Lung	30 (22.7)
		Other*	37 (28.0)
Body surface area (m2)		1–1.49	83 (40.7)
		1.5–1.99	116 (56.9)
		≥ 2	5 (2.5)
Body mass index		< 18.5	84 (41.2)
		18.5–24.99	90 (44.1)
		25–29.99	17 (8.3)
		30 and above	13 (6.4)
ECOG performance status (ECOG-PS)	Cycle 1	0–1	147 (72.1)
		≥ 2	57 (27.9)
	Cycle 2	0–1	149 (73.0)
		≥ 2	55 (26.96)
	Cycle 3	0–1	168 (82.35)
		≥ 2	36 (17.64)
	Cycle 4	0–1	188 (92.2)
		≥ 2	16 (7.84)
	Cycle 5	0–1	190 (93.1)
		≥ 2	14 (6.8)
	Cycle 6	0–1	193 (94.6)
		≥ 2	11 (6.4)

Abbreviations: ADC Adenocarcinoma, HTN Hypertension, DM Diabetic mellitus, RVI Retroviral infection, HF Heart failure, ECOG-PS Eastern Cooperative Oncology Group Performance Status, GIT Gastrointestinal tract. Note: Other* HTN and DM = 2, HTN and RVI = 2, HF = 3, DM and asthma = 2, DM and RVI = 2, hyperthyroidism = 2, and Epilepsy = 1

Treatment-related characteristics of the patients

Most patients (79.9%) received fluoropyrimidine-based chemotherapy, predominantly FOLFOX (73.5%, n = 150); while gemcitabine-cisplatin was used in 8.3% (n = 17). Among the 61 patients (29.9%) receiving second-line chemotherapy, 32 (15.5%) received FOLFIRI and 17 (8.3%) FOLFOX. Regarding dosing, 108 patients (62.7%) received 5-FU and oxaliplatin with an additional 10% bolus of 5-FU. Surgical intervention was performed for 59.3% (n = 121) of patients, while 4.9% (n = 10) received radiotherapy, and 19.6% (n = 40) were scheduled to receive radiation therapy. Standard-dose chemotherapy was administered to 77.0% (n = 157) of patients, whereas 11.8% (n = 24) and 11.3% (n = 23) received lower and higher doses, respectively. Due to the predominance of FOLFOX and the limited sample size, regimen type, dose intensity, and treatment intent were excluded from multivariable analyses (Table 3).

Table 3.

Treatment related characteristics of GI tract cancer patients treated at the oncology unit of TASH and SPHMMC between June 2023 and February 2024, Addis Ababa, Ethiopia (n = 204)

Variables	Frequency n (%)
Fluor-pyrimidine based used at first-line
No	41 (20.1)
FOLFOX	150 (73.5)
Cap-OX	8 (3.9)
5-FU plus Lecuvorin	5 (2.5)
Taxanes/Platinum based used as first-line
No	163 (79.9%)
Gemcitabine Plus Cisplatine	17 (8.3)
Carboplatin plus Paclitaxel	15 (7.4)
Cisplatine Plus Paclitaxel	6 (2.9)
Other*	3 (1.5)
Chemotherapy regimen used as the Second-line
No	143 (70.1)
FOLFOX	17 (8.3)
FOLFIRI	32 (15.7)
Carboplatin plus paclitaxel	4 (2.0)
Cisplatin plus Paclitaxel	2 (1.0)
Other **	6 (2.9)
Targeted therapy used
No	185 (90.7)
Yes	19 (9.3)
Chemotherapy used as
Neo- Adjuvant	64 (31.4)
Adjuvant	38 (18.6)
Palliative	102 (50.0)
Dose of FOLFOX given
Not	41 (20.1)
5-FU, Oxaliplatin plus 10% added 5-FU	128 (62.7)
5-FU, Lecuvorin plus Oxaliplatin	35 (17.2)
Total chemotherapy dose
At standard dose	157 (77.0)
Dose too high	23 (11.3)
Dose too low	24 (11.8)

Abbreviations: FOLFOX Fluorouracil, leucovorin, and oxaliplatin, CAP-OX Capecitabine and oxaliplatin, FLOT Fluorouracil, leucovorin, oxaliplatin, and docetaxel, FOLFOXIRI Fluorouracil, leucovorin, oxaliplatin and irinotecan. Notes: Other*: gemcitabine alone = 1; cisplatin plus etoposide = 2; other** FLOT = 2; Cap-OX = 2; gemcitabine + paclitaxel = 2

Magnitude of chemotherapy-induced toxicity:—hematological and non-hematological toxicities

Among the 204 patients with gastrointestinal (GI) cancer who completed chemotherapy, a total of 1,499 adverse drug reactions (ADRs) were recorded. Hematologic toxicities accounted for 47.4% (n = 710) of all ADRs, while non-hematologic toxicities comprised 52.6% (n = 789). The highest incidence of ADRs occurred during cycle 2 (32.2%, n = 482), followed by cycle 3 (24.4%, n = 366). Within hematologic toxicities, anemia (40.3%, n = 286) and neutropenia (39.0%, n = 277) were the most frequent. Non-hematologic toxicities were led by nausea and vomiting (12.5%, n = 99), followed by fatigue (12.4%, n = 98) and peripheral neuropathy (11.2%, n = 88) (Figs. 2 and 3).

Fig. 2.

Cycle-wise Distribution of Chemotherapy-Related Toxicities at the oncology unit of TASH and SPHMMC between June 2023 and February 2024, Addis Ababa, Ethiopia (n = 204)

Fig. 3.

Frequency of chemotherapy-induced hematological and non-hematological toxicity in gastrointestinal cancer patients treated at the oncology unit of TASH and SPHMMC from June 2023 to February 2024, Addis Ababa, Ethiopia (n = 204). Note: Others *Mucositis = 6, headache = 4, fever = 3, photosensitivity = 3, Rash = 3 Conjunctivitis = 4, hypovolemic shock = 2 urinary incontinence = 2 and tooth discolouration = 2

Chemotherapy-induced toxicity grading by National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE)

Hematological toxicity grading

The severity of hematologic toxicities was graded according to NCI CTCAE criteria. Overall, 556 (37.1%) documented toxicities were grade ≥ 3; of these, 254 (45.7%) were hematologic toxicities. Neutropenia accounted for 79.9% (n = 203) of hematologic toxicities and was most frequently observed during cycle 3 (81 patients, 40.0%), of whom 63(77.8%) had grade ≥ 3 toxicity. Anemia was most common in cycle 2 (80 patients, 39.2%); however, most anemia episodes were grade 1–2 (92.5%, n = 74) (Table 4).

Table 4.

National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 for severity of toxicity assessment of hematological toxicity grading of GIT cancer patients treated at the oncology unit of TASH and SPHMMC between June 2023 and February 2024, Addis Ababa, Ethiopia (n = 204)

Variable n (%)	Chemotherapy Cycle (n = 204)
	1 st cycle	2nd cycle	3rd cycle	4th cycle	5th cycle	6th cycle
Neutropenia	55 (26.9)	66 (32.4)	81 (40.0)	24(11.7)	30 (14.7)	21(10.3)
Grade 1–2	13 (23.6)	13 (19.7)	18 (22.2)	9 (37.5)	12 (40.0)	9 (42.9)
Grade ≥ 3	42 (76.4)	53 (80.3)	63 (77.8)	15 (62.5)	18 (60.0)	12 (57.0)
Anemia	57 (27.9)	80 (39.2)	59 (28.9)	44 (21.6)	32 (15.7)	14 (6.8)
>Grade 1–2	53 (92.9)	74(92.5)	58 (98.3)	44 (100)	32 (100)	14 (100)
Grade ≥ 3	4 (7.1)	6 (8.1)	1 (1.7)	__	__	__
Thrombocytopenia	10 (4.9)	25 (12.3)	16(7.8)	18 (8.8)	9 (4.4)	3 (1.5)
Grade 1–2	8 (80.0)	21 (84.0)	12 (75.0)	17 (94.4)	6 (66.7)	3
Grade ≥ 3	2 (10.0)	4 (16.0)	4 (25.0)	1 (5.5)	3 (33.3)	__
Pancytopenia	1 (0.5)	6 (2.9)	3 (1.5)	3 (1.5)	__	2 (2.7)
Grade 1–2	__	2 (33.3)	__	1	__	__
Grade ≥ 3	1	4 (66.6)	3	2	__	2
Leucopenia	5 (2.5)	19 (9.3)	12 (5.8)	4 (1.9)	9 (4.4)	2 (1.0)
Grade 1–2	4	15 (78.9)	8 (66.7)	3	7 (77.8)	__
Grade ≥ 3	1	4 (21.0)	4 (33.3)	1	2 (22.2)	2

Key note: n = 201 at cycle 6

Non-hematological toxicity grading

Non- hematologic toxicities accounted for 54.3% (n = 302) of documented grade ≥ 3 chemotherapy-induced toxicities. During cycle 2, nausea and vomiting occurred in 31patients (15.2%), of whom 7 (22.6%) experienced grade ≥ 3 toxicity. Alopecia was the most frequent non-hematologic toxicity, affecting 45 patients (22.1%), with 36 (80.0%) experienced grade ≥ 3 toxicity. Diarrhea affected 32 patients (15.7%) during cycle 2, with 4 (12.5%) experiencing grade ≥ 3 toxicity. The full distribution according to cycle and toxicity grade is presented in Table 5.

Table 5.

National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0 for severity of toxicity assessment of non-hematological toxicity grade of GI tract cancer patients treated at the oncology unit of TASH and SPHMMC between June 2023 and February 2024, Addis Ababa, Ethiopia (n = 204)

Variable N (%)	Chemotherapy Cycle (n = 204)
	1 st cycle	2nd cycle	3rd cycle	4th cycle	5th cycle	6th cycle
Nausea/vomiting	33 (16.2)	31 (15.2)	13 (6.4)	10 (4.9)	5 (2.5)	7 (3.4)
Grade 1–2	31 (94.0)	24 (77.4)	6(46.2)	3 (30.0)	1 (20.0)	2 (28.5)
Grade ≥ 3	2 (6.0)	7 (22.6)	7 (53.8)	7 (70.0)	4 (80.0)	5 (71.4)
Diarrhea	18 (8.8)	32 (15.7)	22 (10.8)	3 (1.5)	3 (1.5)	1 (0.5)
Grade 1–2	15 (83.3)	18 (56.2)	5 (22.7)	__	3	__
Grade ≥ 3	3 (16.6)	4 (43.8)	17 (77.3)	3	__	1
Constipation	8 (3.9)	13 (6.4)	16 (7.8)	5 (2.5)	4 (1.9)	3 (1.5)
Grade 1–2	4 (1.9)	11 (5.4)	11 (5.4)	2 (1.0)	__	1 (0.5)
Grade ≥ 3	4 (1.9)	2 (1.0)	5 (2.5)	3 (1.5)	4 (1.9)	2 (1.0)
Abdominal pain	11 (5.4)	5 (2.5)	14 (6.9)	5 (2.5)	1 (0.5)	2 (1.0)
Grade 1–2	10 (4.9)	4 (1.9)	10 (4.9)	5 (2.5)	1 (0.5)	2 (1.0)
Grade ≥ 3	1 (0.5)	1 (0.5)	4 (1.9)	__	__	__
Peripheral Neuropathy	10 (4.9)	31 (15.2)	24 (11.8)	10 (4.9)	8 (3.9)	5 (2.5)
Grade 1–2	9 (4.4)	25 (12.3)	15 (7.4)	7 (3.4)	2 (1.0)	1 (0.5)
Grade ≥ 3	1 (0.5)	6 (2.9)	9 (4.4)	3 (1.5)	6 (2.9)	4 (1.9)
Hand Foot Syndrome	10 (4.9)	20 (9.8)	16 (7.8)	1 (0.5)	2 (1.0)	__
Grade 1–2	9 (4.4)	19 (9.3)	8 (3.9)	1 (0.5)	1 (0.5)	__
Grade ≥ 3	1 (0.5)	1 (0.5)	8 (3.9)	__	1 (0.5)	__
Alopecia	16 (7.8)	45 (22.1)	7 (3.4)	3 (1.5)	2 (1.0)	__
Grade 1–2	3 (18.8)	9 (20.0)	__	1	1 (0.5)	__
Grade ≥ 3	13 (81.2)	36(80.0)	7	2	1 (0.5)	__
Fatigue	18 (8.8)	31 (15.2)	22 (10.8)	13 (6.4)	10 (4.9)	4 (1.9)
Grade 1–2	18 (8.8)	25 (12.3)	18 (8.8)	9 (4.4)	10 (4.9)	3 (1.5)
Grade ≥ 3	_	6 (2.9)	4 (1.9)	4 (1.9)	__	1 (0.5)
Dysphagia	__	4 (1.9)	3 (1.5)	__	__	1 (0.5)
Grade 1–2	__	3 (1.5)	2 (1.0)	__	__	1 (0.5)
Grade ≥ 3	__	1 (0.5)	1 (0.5)	__	__	__
Nail Pigmentation	5 (2,5)	7 (3.4)	3 (1.5)	__	__	__
Grade 1–2	5 (2.5)	7 (3.4)	3 (1.5)	__	__	__
Grade ≥ 3	__	___	__	__	__	__
Dry Cough	7 (3.4)	13 (6.4)	15(7.4)	10 (4.9)	5 (2.5)	1 (0.5)
Grade 1–2	3 (1.5)	8(3.9)	9 (4.4)	4 (1.9)	5 (2.5)	__
Grade ≥ 3	4 (1.9)	5 (2.5)	6 (2.9)	6 (2.9)	__	1 (0.5)
Nephrotoxicity	6 (2.9)	20 (9.8)	21 (10.3)	8(3.9)	5 (2.5)	4 (1.9)
Grade 1–2	5 (2.5)	19 (9.3)	19 (9.3)	8(3.9))	5 (2.5)	4 (1.9)
Grade ≥ 3	1 (0.5)	1 (0.5)	2 (1.0)	__	__	_-
Other*	6 (2.9)	15(7.4)	7 (3.4)	1 (0.5)	__	__
Grade 1–2	5 (2.5)	10 (4.9)	4 (1.9)	__	__	__
Grade ≥ 3	1 (0.5)	5 (2.5)	3 (1.5)	1 (0.5)	__	__

Note: Other* Mucositis, headache, fever, photosensitivity, rash, conjunctivitis, hypovolemic shock, urinary incontinence, and tooth discoloration

Complication frequencies during chemotherapy treatment

Among the 204 patients, grade ≥ 3 complications were most commonly due to febrile neutropenia (24.0%, n = 49), which peaked during cycle 2 (22.6%) and declined progressively to 3.4% by cycle 6. Pneumonia (10.3%, n = 21) and thrombosis/embolism (8.8%, n = 18) were less frequent. Overall, the majority of grade ≥ 3 complications occurred during the initial treatment phase, with 72% arising in cycles 1–3 compared to 28% in cycles 4–6 (Fig. 4).

Fig. 4.

Frequency of Febrile Neutropenia, Thrombosis/Embolism, Sepsis, and Pneumonia across Chemotherapy Cycles at the oncology unit of TASH and SPHMMC between June 2023 and February 2024, Addis Ababa, Ethiopia (n = 204)

Causality and severity assessment of adverse drug reactions

Causality was assessed with the Naranjo Probability Scale. Most adverse drug reactions were classified as “probable” (54.8%, score 5–8), with hematologic toxicities accounting for 54.6%. Adverse drug reactions classified as “possible” accounted for 32.7% (score 1- 4), most of which were non hematologic (60.8%) (Fig. 5).

Fig. 5.

Causality assessment of chemotherapy-induced toxicity in gastrointestinal cancer patients treated at the oncology unit of TASH and SPHMMC from June 2023 to February 2024, Addis Ababa, Ethiopia (n = 204)

Among hematological toxicities, anemia (37.2%) and neutropenia (35.8%) were the most frequent. Non-hematological toxicities were primarily peripheral neuropathy (15.8%), fatigue (15.3%), and alopecia (12.1%). According to the Hartwig and Siegel severity scale, most cumulative toxicities were classified as mild (53.7%) or moderate (42.5%). (Table 6).

Table 6.

Causality assessment of individual toxicity using the Naranjo algorithm and Hartwig severity scale in gastrointestinal cancer patients treated at the oncology unit of TASH and SPHMMC from June 2023 to February 2024, Addis Ababa, Ethiopia (n = 204)

Variables	Number of adverse drug reactions
	Naranjo’s Algorithm					Hartwig’ severity scale
	N	Doubtful	Possible	Probable	Definite	Mild	Moderate	Severe
Cumulative Toxicity	1499	108	490	822	79	806	637	56
Combined Hematological	710	17	192	449	52	274	410	26
Anemia	286	13	80	182	11	132	145	9
Neutropenia	277	2	79	175	21	83	187	7
Thrombocytopenia	81	0	15	56	10	37	40	4
Pancytopenia	15	0	3	5	7	7	7	1
Leucopenia	51	2	15	31	3	15	31	5
Combined Nonhematological	789	91	298	373	27	532	227	30
Nausea/vomiting	99	21	51	24	3	42	51	6
Diarrhea	79	2	49	26	2	34	43	2
Constipation	49	12	19	17	1	28	21	0
Abdominal pain	38	6	13	19	0	32	6	0
Nephrotoxicity	65	3	23	33	6	60	5	0
Peripheral neuropathy	88	3	23	59	3	79	9	0
Hand foot syndrome	49	2	5	38	4	10	0	10
Alopecia	73	2	24	45	2	41	8	0
Nail pigmentation	15	0	6	9	0	15	0	0
Fatigue	98	2	39	57	0	78	21	0
Dry Cough	51	9	14	23	5	33	16	2
Pneumonia	21	11	8	2	0	0	12	9
Thrombosis/embolism	18	4	6	7	1	0	14	4
Sepsis	4	5	0	0	4	0	4	5
Other	5	12	12	0	5	10	17	2

Note: Hartwig’ severity scale: Mild = level 1 and 2, Moderate = level 3 and 4, Severe = 5, 6 and 7. Naranjo’s score: total score ≥ 9: definite. Total Score 5–8:Probable Total Score 1–4: possible Total Score ≤ 0: Doubtful

Chemotherapy-induced toxicity evaluation

Distribution of highest-grade toxicity by patient characteristics

Among the 204 gastrointestinal cancer patients, moderate to severe (Grade ≥ 3) toxicities were frequently observed. At TASH and SPHHMC, the rates of hematological toxicity rates were 62.8% and 60.1%, respectively, whereas non-hematological toxicities occurred in 66.9% and 69.6% of patients, respectively. Female patients experienced slightly higher severe hematological toxicities compared with males (64.8% vs 59.6%). Patients aged ≥ 60 years had the highest rates of toxicities, with hematologic and non-hematologic toxicities reported in 63.6% and 75.8% of patients, respectively.

Obese patients had the highest rate of hematological toxicity (84.6%). Squamous cell carcinoma was associated with a higher rate of severe hematologic toxicity (71.4%).

Laboratory parameters showed that 69.2% of patients with absolute neutrophil counts < 1500 cells/µL experienced severe hematological toxicity, while 34.0% of those with anemia developed severe hematological toxicity. Among treatment-related factors, higher-than-standard chemotherapy doses resulted in severe hematological toxicity in 73.9% of patients.Cisplatin plus paclitaxel combination had the highest toxicity rates (85.7% for both hematologic and non-hematologic toxicity), while FOLFOX, showed 64.0% (hematologic) and 69.0% (non-hematologic). Patients with comorbidity (76.9%) and metastasis (71.2%) had higher rates of sever non-hematologic toxicity. By cancer site, esophageal cancer had the highest rate of severe hematologic toxicity (74.3%), while gastric and colon cancers had the highest non-hematologic toxicity rates (75.0% and 72.3%, respectively). The fully stratified results are presented in Table 7.

Table 7.

Distribution of highest grade hematological and non-hematological toxicities by patient characteristics among GIT cancer patients treated at TASH and SPHMMC, Addis Ababa, Ethiopia (n = 204)

		Combined hematological Toxicity		Combined Non-hematological Toxicity
Variables	n	Low/Mild	Moderate to Severe	Low/Mild	Moderate to Severe
Study Settings
TASH	148	55 (37.2)	93 (62.8)	49 (33.1)	99 (66.9)
SPHMMC	56	22 (39.2)	34 (60.1)	17 (30.4)	39 (69.6)
Gender
Male	99	40 (40.4)	59 (59.6)	32 (32.3)	67 (67.7)
Female	105	37 (35.2)	68 (64.8)	34 (32.4)	71 (67.6)
Age (years)
< 40	37	13 (35.1)	24 (64.9)	13 (35.1)	24 (64.9)
40–49	43	22 (51.2)	21 (48.8)	13 (30.2)	30 (69.8)
50–59	58	29 (50.0)	29 (50.0)	24 (41.4)	34(58.6)
≥ 60	66	24 (36.4)	42 (63.6)	16 (24.2)	50 (75.8)
Body Weight
< 18.5	84	27 (32.1)	57 (67.8)	27 (32.1)	57 (67.8)
18.5–24.99	90	38 (42.2)	52 (57.8)	34 (37.7)	56 (62.2)
25–29.99	17	6 (35.3)	11 (64.7)	14 (82.4)	3 (17.6)
30 and above	13	2 (15.4)	11 (84.6)	2 (15.4)	11 (84.6)
Education Status
No formal education	95	36 (37.9)	59 (62.1)	32 (33.7)	63 (66.3)
Primary school	41	15 (36.6)	26 (63.4)	10 (24.4)	31 (75.6)
Secondary school	31	11 (35.5)	20 (64.5)	11 (35.5)	20 (64.5)
Higher education	37	15 (40.5)	22 (59.5)	13 (35.1)	24 (64.8)
Histopathology
ADC	190	73 (38.4)	117(61.6)	63 (33.2)	127(66.8)
SCC	14	4 (28.6)	10 (71.4)	3 (21.4)	11 (78.6)
Pathological grading
WD	147	62 (42.2)	85(57.8)	45 (30.6)	102(69.4)
MD	11	3 (27.7)	8 (72.7)	2(18.2)	9 (81.8)
PD	46	12(26.1)	34 (73.9)	19 (41.3)	27 (48.7)
Cancer Site
Rectal cancer	51	19(37.3)	32(62.7)	15 (29.4)	36(70.6)
Colon cancer	47	21(44.7)	26(55.3)	13 (27.7)	34(72.3)
Gastric cancer	28	10 (35.7)	18(64.3)	7(25.0)	21(75.0)
Esophageal	35	9(25.7)	26(74.3)	15 (42.8)	20(57.1)
Other*	43	18(41.8)	25(58.1)	16(37.2)	27(62.8)
Stage of cancer
Stage III	72	24 (33.3)	48 (66.7)	28 (38.9)	44 (61.1)
Stage IV	132	53 (40.1)	79 (59.9)	38 (28.8)	94 (71.2)
Comorbidity
No	139	56 (40.3)	83 (59.7)	51 (36.0)	88(64.0)
Yes	65	32 (41.5)	33 (50.8)	15 (24.6)	50 (75.4)
Metastasis
No	72	24 (33.3)	48 (66.7)	28 (38.9)	44 (61.1)
Yes	132	53 (40.1)	79 (59.9)	38 (28.8)	94 (71.2)
Site of metastasis
No met	72	24(33.3)	48 (66.7)	28 (38.9)	44 (61.1)
Lung	18	7(38.9)	11(61.1)	6(33.3)	12 (66.7)
Liver	47	24(51.1)	23(48.9)	14 (29.8)	33 (70.2)
Liver and Lung	30	10 (33.3)	20 (66.7)	5(16.7)	25 (83.3)
Other site	37	12 (32.4)	25 (67.6)	13(35.1)	24 (64.9)
ECOG PS 0–1	94	45 (47.9)	49 (52.1)	37(39.4)	57(60.6)
≥ 2	110	43 (39.1)	67 (60.5)	29(26.4)	81(73.6)
White blood cell
< 3000 cell/nL	5	1 (20.0)	4 (80.0)	0 (0.0)	5 (100.00
≥ 3000 cell/nL	199	76 (38.2)	123(62.8)	66 (33.2)	133(66.8)
Hemoglobin (gm/dL)
< 10.0 F, < 11.0 M	47	31 (66.0)	16 (34.0)	19 (40.4)	28 (59.6)
≥ 10.0 F, ≥ 11.0 M	157	57 (36.3)	100 (63.7)	47 (29.9)	110(70.1)
Absolute neutrophil count
< 1500	52	16 (30.8)	36 (69.2)	21 (37.5)	35 (62.5)
≥ 1500	152	72 (47.4)	80 (52.6)	45 (30.4)	103(69.6)
Platelet (cell/dl)
< 75,000	7	3 (42.9)	4(57.1)	3 (42.9)	4 (57.1)
≥ 75,000	197	74 (37.6)	123(62.4)	63 (32.0)	134(68.0)
Fluorpyrimidine Used
No	41	26(63.4)	15(36.6)	19(46.3)	22(53.7)
Yes	163	51(31.3)	112(68.7)	47(28.8)	116(71.2)
Fluorpyrimidine
FOLFOX	150	54(36.0)	96(64.0)	46(30.6)	104(69.0)
Cap-OX	8	4(50.0)	4(50.0)	2(25.0)	6(75.0)
5-FU plus Leucovorin	5	0	5	1	4
Taxanes/Platinum
GC*	17	11(64.7)	5(35.3)	5(29.4)	12(70.5)
PC*	15	7(46.7)	8(53.3)	10(66.7)	5(33.3)
CP*	7	1(14.3)	6(85.7)	1(14.3)	6(85.7)
FOLFOX Given as
5-FU.Oxaliplatin plus 10% added 5-FU	129	44 (34.1)	85 (65.8)	35(27.1)	94(72.8)
5-FU,Lecuvorin plus Oxaliplatin	34	14 (41.2)	20 (58.8)	13(38.2)	21(61.7)
Chemotherapy dose
At Standard dose	157	61 (45.2)	86 (54.8)	51 (32.5)	106 (67.5)
Dose To High	23	6 (26.1)	17 (73.9)	9 (39.1)	14 (60.9)
Dose To Low	24	11(45.8)	13 (54.2)	6(25.0)	18(75.0)

Abbreviations: ECOG PS Eastern Cooperative Oncology Group performance status. BMI Body mass index, ADC Adenocarcinoma, WD Well differentiated, MD Moderately differentiated, PD Poorly differentiated, SCC Squamous Cell Carcinomas, GC* Gemcitabine plus Cisplatin, PC* Paclitaxel plus Carboplatin, CP* Cisplatin plus Paclitaxel, F-Female, M Male, Low/Mild Grade 1–2, Moderate to Severe-Grade 3–5,

Chemotherapy regimen distribution by cancer site

Colorectal cancers were mainly treated with FOLFOX (rectal cancer 88.2%, colon cancer 93.6%), while esophageal (57.1%) and other cancers (44.2%) more often received non-fluoropyrimidine regimens. Among the 41 patients (20.1%) who didn’t receive fluoropyrimidines, gemcitabine–cisplatin was most commonly used in “other” cancers (39.0%), whereas carboplatin–paclitaxel was frequently used in esophageal cancer (24.4%). Overall, 76.9% of patients received standard dose chemotherapy; dose reductions were most frequent among patient with rectal cancer (21.6%). The majority of FOLFOX regimens were administered in an intensified form consisting of 5-FU plus oxaliplatin with 10% dose escalation (62.3%) (Table 8).

Table 8.

Distribution of Fluoropyrimidine-Based Regimens, FOLFOX Variants, Non-Fluoropyrimidine Therapies, and Chemotherapy Dosing by Primary Tumor Site at TASH and SPHMMC, June 2023–February 2024 (n = 204)

Category	Variable	Total n(%)	Rectal ca	Colon ca	Gastric ca	Esophageal ca	Other
Fluor-pyrimidine based used	No	41 (20.1%)	-	2 (4.3%)	4 (14.3%)	14 (40.0%)	21 (48.8%)
	Yes	163 (79.9%)	51 (100%)	45 (95.7%)	24 (85.7%)	21 (60.0%)	22 (51.2%)
Fluor-pyrimidine based Types	Not Used	41 (20.1%)	-	2 (4.3%)	4 (14.3%)	14 (40.0%)	21 (48.8%)
	FOLFOX	150 (73.5%)	45 (88.2%)	44 (93.6%)	22 (78.6%)	20 (57.1%)	19 (44.2%)
	Cap OX	8 (3.9%)	6 (11.8%)	1 (2.1%)	-	1 (2.9%)	-
	5-FU + Leucovorin	5 (2.5%)	-	-	2 (7.1%)	-	3 (7.0%)
FOLFOX Regimen Given	Not given	43 (21.1%)	5 (9.8%)	2 (4.3%)	5 (17.9%)	9 (25.7%)	22 (51.2%)
	5FU + Leuc + Ox	34 (16.7%)	10 (19.6%)	9 (19.1%)	4 (14.3%)	7 (20.0%)	4 (9.3%)
	5FU + Ox + 10%Add	127 (62.3%)	36 (70.6%)	36 (76.6%)	19 (67.9%)	19 (54.3%)	17 (39.5%)
Taxanes/Platinum	Cisplatin + Paclitaxel	6 (14.6%)	-	-	1 (2.4%)	3 (7.3%)	2 (4.7%)
	Gemcitabine + Cisplatin	17 (41.5%)	-	-	1 (2.4%)	-	16 (39.0%)
	Carboplatin + Paclitaxel	15 (36.6%)	-	-	2 (4.9%)	10 (24.4%)	3 (7.3%)
	Gemcitabine alone	1 (2.4%)	-	-		1 (2.4%)	-
	Cisplatin + Etoposide	2 (4.9%)	-	2 (4.9%)	-	-	-
Total chemotherapy dose	Standard	159 (77.9%)	37 (72.5%)	35 (74.5%)	22 (78.6%)	30 (85.7%)	35 (81.4%)
	Too high	21 (10.3%)	3 (5.9%)	8 (17.0%)	2 (7.1%)	3 (8.6%)	5 (11.6%)
	Too low	24 (11.8%)	11 (21.6%)	4 (8.5%)	4 (14.3%)	2 (5.7%)	3 (7.0%)

Abbreviations: FOLFOX Fluorouracil, leucovorin, and oxaliplatin, CAP-OX Capecitabine and oxaliplatin

Notes: Other*: gemcitabine alone = 1; cisplatin plus etoposide = 2; other** FLOT = 2; Cap-OX = 2; gemcitabine + paclitaxel = 2

A total of 753 supportive interventions were administered, including G-CSF (n = 253, 33.6%), multivitamins (n-98,13.0%), antimicrobials (n = 86, 11.4%), and blood transfusions (n = 77, 10.2%). Overall, 53.4% of patients received G-CSF at least once, with 50 patients (24.5%) receiving it during cycle 2. Antimicrobials were administered to 85 patients (41.75) at least once. Supportive care utilization according to treatment cycle and cumulative totals is presented in Table 9.

Table 9.

Use of G-CSF, antibiotics, antiemetics, and other supportive care interventions by chemotherapy cycle among GIT cancer patients at the oncology units of TASH and SPHMMC, Addis Ababa, Ethiopia (June 2023–February 2024, n = 204)

Variable N (%)	Chemotherapy cycle						supportive measure
	1 st cycle	2nd cycle	3rd cycle	4th cycle	5th cycle	6thcycle
G-CSF Use	17 (8.3)	50(24.5)	37(18.1)	22 (10.8)	24(11.8)	3 (1.5)	253 (33.6)
Anti-Emetics	6 (2.9)	17 (8.3)	7 (3.4)	5 (2.5)	5 (2.5)	4 (2.0)	44 (5.8)
Blood Transfused	14 (6.9)	17 (8.3)	17 (8.3)	16 (7.8)	9 (4.4)	4 (2.0)	77(10.2)
Multi-Vitamins	21(10.3)	24(11.8)	17 (8.3)	14 (6.9)	14 (6.9)	8 (3.9)	98 (13.0)
Anti-Microbial	14 (6.9)	33(16.2)	22(10.8)	8 (3.9)	4 (2.0)	5 (2.5)	86 (11.4)
Laxative	2 (1.0)	10 (4.9)	7 (3.4)	3 (1.5)	0	3 (1.5)	25 (3.3)
Anti-Diarrhea	5 (2.5)	10 (4.9)	12 (5.9)	6 (2.9)	2 (1.0)	1 (1.0)	36 (4.8)
Analgesics	18 (8.8)	16 (7.8)	13 (6.4)	3 (1.5)	6 (2.9)	4 (2.0)	60 (7.9)
Atropine	3 (1.5)	8 (3.9)	3 (1.5)	2 (1.0)	2 (1.0)	__	18(2.4)
Diphenhydramine	1 (1.0)	3 (1.5)	6(3.0)	4 (2.0)	2 (1.0)	__	16 (2.1)
Amitriptyline	__	__	2 (1.0)	__	2 (1.0)	1 (1.0)	5(0.7)
Anti-coagulant	3 (1.5)	4 (2.0)	5 (2.5)	__	__	__	12(1.6)
Pregabaline	__	2 (1.0)	2 (1.0)	__	__	__	4 (0.5)
Anti-viral/Anti-fungal	__	2 (1.0)	2 (1.0)	__	__	__	4 (0.5
Other *	4 (2.0)	5 (2.5)	2 (1.0)	3 (1.5)	1 (1.0)	__	15 (1.9)

Abbreviations: *G-CSF Granulocyte colony stimulating factor, GIT Gastrointestinal tract

Note. Percentages in the cycle-specific columns represent the number of patients receiving each intervention in that cycle divided by the total number of patients (n = 204). The Total column shows the cumulative number of administrations (events) across all cycles and the corresponding percentage of total patient-cycle exposures

Association of highest grade of adverse drug reaction with patient’s characteristics

Variables with a p-value ≤ 0.25 in the bivariate logistic regression analysis were included in the multivariable logistic regression model. Multicollinearity among independent variables was assessed using collinearity diagnostics; the largest condition index was 6.528 (< 10).. The model demonstrated good fit (Hosmer–Lemeshow χ² = 9.265, df = 8, p = 0.320), explained a moderate proportion of the variance (Cox & Snell R² = 0.184; Nagelkerke R² = 0.247), and was statistically significant overall (Omnibus Test χ² = 41.540, df = 12, p < 0.001) (Supplementary Tables).

In the multivariable logistic regression analysis, three variables were significantly associated with high grade of ADRs, while one variable showed a significant protective effect. Pathological grade, low baseline hemoglobin and low baseline absolute neutrophil count were significant associated with high grade ADRs. Patients with poorly differentiated tumors had higher odds of experiencing moderate to severe ADRs compared with those with well-differentiated tumors (AOR = 4.410, 95% CI: 1.823–10.666, p = 0.001). Similarly, low baseline hemoglobin (< 10.0 g/dL for females and < 11.0 g/dL for males) was associated with higher odds of severe ADRs (AOR = 4.456, 95% CI: 2.021–9.824, p = 0.001), while a baseline absolute neutrophil count < 1500 cells/µL was associated with higher odds of severe ADRs (AOR = 2.237, 95% CI: 1.061–4.717, p = 0.034). In contrast, G-CSF use was independently associated with a lower odd of high-grade ADRs (AOR = 0.508, 95% CI: 0.260–0.995, p = 0.048) (Table 10).

Table 10.

Bivariable and multivariable logistic regression analyses of factors associated with moderate to severe adverse drug reactions (NCI CTCAE version 5.0) among gastrointestinal cancer patients treated at the Oncology Unit of Tikur Anbessa Specialized Hospital, Addis Ababa, Ethiopia, June 2023–February 2024 (n = 204)

Variables	Low/Mild	Moderate to Severe	COR(95% C.I)	AOR (95%C.I)	P-value
Age (Year)
< 40 (Ref)	13 (35.1)	24 (64.9)	1	1
40–49	22 (51.2)	21 (48.8)	0.517 (0.210–1.274)	0.480 (0.172–1.336)	0.168
50–59	29 (50.0)	29 (50.0)	0.542 (0.232–1.266)	0.428 (0.165–1.112)	0.081
60 and Above	24 (36.4)	42 (63.6)	0.948 (0.409–2.198)	0.894 (0.344–2.327)	0.819
ECOG-PS
ECOG 0–1 (Ref)	45 (47.9)	49 (52.1)	1	1
ECOG ≥ 2	43 (39.1)	67 (60.9)	1.431 (0.820–2.498)	1.453 (0.751–2.810)	0.268
Pathological Grading
Well differentiated (Ref)	72 (49.0)	75 (51.0)	1	1
Moderate differentiated	3 (27.3)	8 (72.7)	2.560 (0.653–10.032)	3.119 (0.715–13.596)	0.130
Poor differentiated	13 (28.3	33 (71.7)	2.437 (1.188–5.000)	4.410 (1.823–10.666)	0.001*
Comorbidity
No (Ref)	56 (40.3)	83 (59.7)	1	1
Yes	32 (49.2)	33 (50.8)	0.696 (0.385–1.259)	1.341 (0.674–2.672)	0.403
Total chemotherapy dose
At standard dose (Ref)	71 (45.2)	86 (54.8)	1	1
Dose to high	6 (26.1)	17 (73.9)	2.339 (0.876–6.247)	1.213 (0.388–3.792)	0.739
Dose to low	11 (45.8)	13 (54.2)	0.976 (0.412–2.311)	0.804 (0.305–2.117)	0.658
Hemoglobin (g/dL)
≥ 10.0 F, ≥ 11.0 M (Ref)	57(36.3)	100 (63.7)	1	1
< 10.0 F, < 11.0 M	31 (66.0)	16 (34.0)	0.294 (0.148–0.584)	4.456 (2.021–9.824)	0.001*
Absolute neutrophil count
≥ 1500(Ref)	72 (47.4)	80 (52.6)	1	1	0.034*
< 1500	16 (30.8)	36 (69.2)	2.025 (1.037–3.955)	2.237 (1.061–4.717)
G-CSF Injection used
No (Ref)	46 (48.4)	49 (51.6)	1	1
Yes	42 (38.5)	67 (61.5)	1.498 (0.858–2.615)	0.508 (0.260–0.995)	0.048*

^*Statistically significant in the multivariable logistic regression analysis (p < 0.05).. COR Crude odds ratio, AOR Adjusted odds ratio, CI Confidence interval, M Male, F Female, G-CSF Granulocyte colony-stimulating factor, ECOG Eastern Cooperative Oncology group, Ref Referenc

Discussion

This study provides a comprehensive analysis of 204 patients with GIT cancer treated at two tertiary care hospitals in Addis Ababa, Ethiopia, and documented a substantial burden of chemotherapy related ADRs. Chemotherapy induced toxicities negatively affect quality of life and treatment continuity, therefore, systematic documentation and reporting are essential to optimize patient management and improve clinical outcomes [26, 27].

The incidence and severity of chemotherapy-induced toxicities observed in this study were generally consistent with findings reported from Nepal, Ireland, Morocco, and France, although some differences were noted [28–31]. We identified 1499 ADRs, with a slightly higher proportion of non-hematologic than hematologic toxicities. The highest incidence of ADRs was observed during cycle 2. This early peak likely reflects initial treatment exposure and dose intensity rather than cumulative toxicity. The pattern of toxicities across treatment cycle further suggests that the early chemotherapy represents a critical period requiring close clinical monitoring.

Eastern Cooperative Oncology Group performance status (ECOG PS) improved across chemotherapy cycles, with the proportion of patients with ECOG PS 0–1 increasing from 72.1% in cycle 1 to 94.6% by cycle 6, whereas the proportion with ECOG PS ≥ 2 decreased from 27.9% to 6.4%. Although this finding may appear unexpected in a predominantly metastatic population, it likely reflects treatment response and enhanced supportive care, including improved symptom control, nutritional support, and proactive management of early toxicities. Similar findings have been reported in chemotherapy-based cohorts, in which patients who continue treatment beyond the early cycles often maintained or improved their performance status, consistent with disease control and optimized supportive care [21, 32]. However, selection bias may have contributed to this observation, as patients with poor treatment tolerance or disease progression were likely to remain into later treatment cycles.

Among patients in this study, fluoropyrimidine-based chemotherapy was the predominant treatment approach, with FOLFOX being the most commonly used regimen (92.0%). This proportion is substantially higher that reported previously in Ethiopia at Tikur Anbessa Specialized Hospital (55.6%) [1] and in studies conducted in Germany, India, and Egypt [6, 33, 34]. These regional variations likely reflect differences in drug availability, healthcare infrastructure, treatment costs, clinical practice patterns, and patient demographics, all of which may influence the consistency of cancer care across settings.

The toxicity profile observed in this study was consistent with the expected effect of cytotoxic chemotherapy, with hematologic toxicities largely reflecting chemotherapy induced myelosuppression [3].Non-hematologic toxicities, particularly nausea, vomiting and fatigue, were among the most frequently reported treatment related adverse effects [31]. The observed toxicity burden may partly reflect limited access to newer antiemetic agents and hematopoietic growth factors in Ethiopia. Furthermore, regional differences in toxicity patterns may be attributable to variations in treatment regimens, patient characteristics, supportive care practices, and adverse-event reporting systems.

According to the NCI CTCAE, 37.1% of all toxicities were grade ≥ 3, with hematologic and non-hematologic toxicities accounting for 45.7% and 54.3%, respectively. Among severe hematologic toxicities, neutropenia was the most frequently reported event (79.9%), whereas nausea and vomiting were the predominant severe non-hematologic toxicities. Compared with a previous study conducted in Gondar, Ethiopia, our findings indicate a higher burden of severe toxicities [3]. In contrast, studies from Bangladesh, Vietnam, Italy, and France reported toxicity patterns more comparable to those observed in our cohort. [31, 35–37]. Collectively, these findings highlight the importance of enhanced toxicity surveillance and individualized supportive care strategies to reduce treatment interruptions and improve patient safety.

Grade ≥ 3 toxicities occurred predominantly during the early treatment period, with 72% observed in Cycles 1–3. This observation is consistent with findings from Western FOLFOX trials, in which 50–75% of neutropenic events occur during the first treatment cycle before dose modification is implemented [38].

Febrile neutropenia was the most common severe toxicity (24%), occurring more frequently than reported in high-resource settings (10–15%) but remaining comparable to rates observed in cohorts from low- and middle-income countries (20–30%) [39]. This finding may reflect the limited use of primary G-CSF prophylaxis and the increased vulnerability of patients during the initial treatment cycles. The subsequent decline in toxicity rates likely reflects dose modifications and supportive care interventions. Nevertheless, the persistence of clinically significant toxicities suggests that early preventive measures remain insufficient, underscoring the need for intensive monitoring during the initial treatment cycles, as recommended by ASCO guidelines. [40]. Antiemetic prophylaxis typically consisted of a 5-HT3 receptor antagonist combined with dexamethasone, whereas NK1 receptor antagonists were rarely used because of cost and limited availability, in contrast to the triple-agent antiemetic regimens commonly recommended and implemented in high-resource settings [11]. Although documented antiemetic utilization appeared low, this may underestimate actual clinical practice because supportive medications are not always consistently recorded in patient charts. Despite these constraints, nausea and vomiting were predominantly grade 1–2, indicating that acceptable symptom control was achieved for most patients [38].

Using the Naranjo probability scale, 54.8% of ADRs in this study were classified as “probable," with hematological reactions accounting for the majority (54.6%), while 32.7% were classified as “possible," predominantly comprising non-hematological reactions (60.8%). According to the Hartwig – Siegel severity scale, 53.7% of ADRs were classified as mild and 42.5% as moderate.Compared with a study conducted in Gondar, Ethiopia, which reported a higher proportion of probable ADRs (68.8%) and a lower proportion of possible ADRs (31.4%) [3], the distribution observed in our study was more comparable to findings from Nepal, where probable and possible ADRs accounted for 47.1% and 46.1% of reactions, respectively, and mild and moderate ADRs accounted for 54.9% and 40.2% [28]. The predominance of mild-to-moderate ADRs suggests that most reactions were manageable and did not necessitate treatment discontinuation or substantial therapeutic modification.The identification of patients at increased risk of chemotherapy-induced toxicity based on demographic and clinical characteristics remains an important component of individualized cancer care [15].

In this study, the incidence of both hematological and non-hematological toxicities was generally comparable between male and female patients, unlike reports from India, where ADRs were more frequently observed in females [41]. Such discrepancies may reflect differences in patient demographics, treatment patterns, genetic susceptibility, and pharmacokinetic variability across study populations [42, 43].

Older patients aged 60 years and above experienced higher rates of moderate to severe combined hematological (63.6%) and non-hematological toxicities (75.8.), consistent with previous reports [28, 42]. The increased risk may reflect age-related reductions in metabolic and excretory capacity, resulting in greater drug exposure and susceptibility to ADRs [16]. These findings support individualized treatment approaches, appropriate dose optimization, and further pharmacokinetic research in older adults.

Our findings further suggest that nutritional status and body composition may influence treatment tolerance. Although evidence regarding obesity remains inconsistent, altered pharmacokinetics and dosing practices may contribute to increased toxicity at BMI extremes. These observations support routine nutritional assessment, including the Patient-Generated Subjective Global Assessment (PG-SGA), and individualized supportive care.

Grade ≥ 3 hematological toxicity was most common among obese (84.6%) and underweight patients (67.8%), suggesting that BMI extremes may increase susceptibility to chemotherapy induced myelosuppression. This findings is consistent with Sucher et al. (2024), who reported greater FOLFOX -related toxicity among patients at BMI extremes, particularly those who were underweight [44]. Similarly, Mao et al. (2017) reported an increased risk of neutropenia among underweight Asian patients, whereas Deng et al. (2024) demonstrated altered oxaliplatin pharmacokinetics in obese individuals [44, 45]. Malnutrition may exacerbate toxicity through hypoalbuminemia and reduced bone marrow reserve; however, BMI was not independently associated with toxicity in multivariable analysis. Future studies should incorporate standardized nutritional assessments and evaluate BMI-stratified dosing strategies in accordance with ASCO recommendations.

In this study, most patients received fluoropyrimidine-based regimens, among whom 68.7% developed grade ≥ 3 hematological toxicities and 71.2% experienced grade ≥ 3 non-hematological toxicities. Among patients treated with FOLFOX, the corresponding rates were 64.0% and 69.0%, respectively. These rates exceed those reported in most clinical trials and real-world studies of FOLFOX-based therapy, where grade 3–4 hematologic toxicities are generally lower or more variable [46–48] Consistent with European Society for Medical Oncology (ESMO) guidelines, grade ≥ 3 neutropenia occurs in approximately 30–50% of patients receiving FOLFOX, while the National Comprehensive Cancer Network (NCCN) categorizes FOLFOX as an intermediate-risk regimen for febrile neutropenia (10–20%). Likewise, the MOSAIC trial reported an overall grade ≥ 3 hematologic toxicity rate of approximately 40% [11, 49, 50]. The higher toxicity burden observed in our cohort may reflect constraints in supportive care, including inconsistent use of G-CSF prophylaxis, limited access to antiemetics, delayed dose modifications, and challenges in outpatient monitoring within Ethiopian oncology practice.

Neuropathy was observed primarily among patients receiving oxaliplatin-based therapy, consistent with the well-established cumulative and dose-dependent neurotoxicity of oxaliplatin. The underlying mechanisms include injury to dorsal root ganglion neurons, mediated by mitochondrial dysfunction, oxidative stress, and ion-channel dysregulation [51–53]. The relatively high frequency observed in our cohort may reflect delayed recognition of cumulative toxicity and limited opportunities for timely dose modification.

Patients with comorbidities had higher rates of grade ≥ 3 hematological toxicity (50.8%), comparable with Li et al. (55.2%) [54], and previous meta-analysis [55]. They also had a higher prevalence of severe non-hematological toxicity than patients without comorbidities (75.4% vs. 64.0%). Similarly, patients with metastatic disease experienced greater toxicity than those without metastasis (71.2% vs. 61.1%) [55]. These findings may be explained by impaired physiological reserve and reduced tolerance to chemotherapy among patients with greater disease burden and coexisting medical conditions.

Colorectal cancer accounted for the majority of cases (74.9%), and FOLFOX was the predominant treatment regimen, consistent with contemporary treatment guidelines. Nevertheless, the observed toxicity burden exceeded that reported in the MOSAIC and OPTIMOX trials [56]. Likely reflecting differences in patient characteristics, supportive care infrastructure, and real-world treatment practices. Patients with esophageal cancer received more heterogeneous treatment regimens, including fluoropyrimidine-based combinations and carboplatin–paclitaxel according to the CROSS protocol, and experienced higher toxicity rates than those reported in the original CROSS trial [57]. Similarly, hepatobiliary malignancies treated predominantly with gemcitabine–cisplatin demonstrated the highest burden of non-hematological toxicity [58]. Collectively, these findings suggest that both regimen intensity and resource-related constraints may contribute to the higher toxicity burden observed in routine clinical practice.

Treatment delays were common, affecting 81.9% of patients, with 48.5% experiencing two or more delays during six treatment cycles. These rates were substantially higher than those reported in Canada, the United State, France and South Korea [59–62]. This finding highlights the need for earlier toxicity recognition, proactive supportive care, and individualized dose optimization to improve treatment continuity and patient safety.

Poorly differentiated tumors were independently associated with severe ADRs (AOR 4.410, p = 0.001), supporting previous evidence that tumor biology may influence treatment-related toxicity [63]. Although prior studies link comorbidities with increased ADRs risk [64], no independent association was observed in our analysis (AOR = 1.341, p = 0.403).

Low baseline hemoglobin was associated with a 4.5-fold increased risk of severe ADRs, consistent with previous reports [65], Similarly, a low baseline absolute neutrophil count independently predicted severe ADRs (AOR = 2.237, p = 0.034), in agreement with earlier studies [66]. These findings underscore the importance of baseline hematologic assessment and risk-stratified toxicity monitoring. In addition, granulocyte colony-stimulating factor (G-CSF) use was independently associated with reduced odds of high-grade ADRs (AOR = 0.508, 95% CI: 0.260–0.995, p = 0.048), supporting its protective role against severe treatment-related toxicities [67].

These findings have important implications for oncology practice in Ethiopia. Routine baseline hematologic evaluation, nutritional assessment, and structured toxicity monitoring should be integrated into the care of patients receiving fluoropyrimidine- and oxaliplatin-based chemotherapy. Strengthening clinical pharmacist involvement, timely dose modification, and appropriate use of supportive measures such as G-CSF may help reduce severe toxicities and improve treatment continuity in resource-limited settings.

Future studies should evaluate the impact of toxicity and dose modification on survival outcomes and investigate pharmacogenetic determinants of fluoropyrimidine- and oxaliplatin-related toxicity. Interventional studies assessing structured toxicity monitoring, pharmacist-led care, and integrated supportive strategies are also needed to strengthen the evidence base for improving chemotherapy safety in Ethiopia.

Strengths and limitations of the study

Strengths

This study has several notable strengths. Its prospective design allowed systematic collection of both patient-reported and clinically documented toxicities throughout six treatment cycles, enhancing the real-world applicability of the findings. The longitudinal follow-up enabled evaluation of toxicity patterns and severity over time. Additionally, the inclusion of two major national referral centers (TASH and SPHMMC) improved the representativeness of patients receiving tertiary oncology care in Ethiopia.

Limitations

Long-term outcomes, including overall and disease-free survival, were not evaluated. The relatively short recruitment period and modest sample size may limit the generalizability of the findings. Health-related quality of life and late-onset toxicities were not assessed. Because participants were recruited consecutively from tertiary referral hospitals, patients with advanced disease may have been over represented, limiting applicability to lower-level healthcare settings and earlier-stage cancers. Subjective toxicities, including fatigue, neuropathy, and gastrointestinal symptoms, may have been under reported, potentially resulting in an underestimation of their true burden. In addition, the absence of regimen-stratified multivariable analyses may have introduced residual confounding, as different chemotherapy regimens have distinct toxicity profiles. Although toxicity patterns were described according to treatment regimen, the independent effects of regimen-specific factors could not be fully evaluated. Future studies should incorporate regimen-specific analyses to better characterize treatment-related toxicity risks.

Conclusion

This prospective study demonstrated a substantial burden of chemotherapy-induced toxicities among patients with gastrointestinal tract cancers receiving treatment at two tertiary hospitals in Ethiopia. Hematological and non-hematological adverse drug reactions were common, with a considerable proportion of patients experiencing severe (grade ≥ 3) toxicities. Neutropenia was the predominant severe hematological toxicity, while nausea and vomiting were the most frequent severe non-hematological toxicities. Poor tumor differentiation, low baseline hemoglobin levels, and low absolute neutrophil counts were independently associated with an increased risk of severe toxicity, whereas G-CSF use was associated with a reduced risk of high-grade adverse drug reactions. These findings highlight the need for systematic toxicity monitoring and risk-based supportive care to improve the safety and continuity of chemotherapy in Ethiopian oncology practice.

Abbreviations

AAU

Addis Ababa University

AOR

Adjusted odds ratio

COR

Crud odds ratio

CRC

Colorectal cancer

ECOG

Eastern Cooperative Oncology Group

5-FU

5-Fluorouracil

ADR

Adverse drug reaction

ASDR

Age- Standardized death rate

CAPEOX or CAPOX

Capecitabine and oxaliplatin

GI

Gastrointestinal

FOLFIRI

5-Fluorouracil, leucovorin and irinotecan

FOLFOX

5-Fluorouracil, leucovorin and oxaliplatin

GBD

Global Burden of Diseases

GCSF

Granulocyte Colony Stimulating Factor

NCCN

National Comprehensive Cancer Network

SPHMMC

St. Paul’s Hospital Millennium Medical College

SPSS

Statistical Package for the Social Sciences

TASH

Tikur Anbessa Specialized Hospital

VEGF

Vascular endothelial growth factor

GLOBOCANO

Global Cancer Statistics

WHO

World Health Organization

Authors’ contributions

S.F.M. conceived and conducted the study, analyzed and interpreted the data, and drafted the manuscript. A.M.F.contributed to the study design and oversaw the research activities. A.B.B. and M.A.W. contributed to the study design and data analysis, supervised the research process, and participated in drafting, critically reviewing, and revising the manuscript. All authors read and approved the final version of the manuscript for publication.

Funding

The study was done with the financial support of Addis Ababa University.

Data availability

All relevant data are included in the article. Extra data are accessed upon reasonable request of the corresponding author.

Declarations

Ethics approval and consent to participate

Ethical approval for the study was obtained from the School of Pharmacy, College of Health Sciences, AAU, ethical review board with Ref no: ERB/SOP/522/15/2023 and also from the Institution Review Board (IRB) of SPHMMC with Ref no. PM 23/212. Before data collection, written permission was obtained from the oncology units of TASH and SPHMMC. This study adhered to the Declaration of Helsinki. The aims of the study were clearly explained to the participants, and information was collected after obtaining written informed consent from each patient or their family. Patients were given the right to refuse or discontinue participation at any time and to ask questions about the study objective. To protect privacy, patients’ names were not used during data collection, personal information was kept confidential, and confidentiality was maintained throughout the study period.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.