INTRODUCTION

In Peru, as in other countries during the most critical periods of the SARS-CoV-2 pandemic, pharmaceutical products were widely used without sufficient scientific evidence on their benefits and harms, including hydroxychloroquine (HCQ), chloroquine, azithromycin (AZI), ivermectin (IVM), tocilizumab (TOB), among others ^1-3. In fact, some media and public figures promoted the inappropriate off-label use of these drugs, which led to self-medication and overmedication; and with it, increased risk of serious adverse drug reactions (ADRs) ⁴.

Although the safety profile of these drugs is known for their approved clinical indications, off-label use for the prevention and/or treatment of SARS-CoV-2 infection could alter their safety profile. Off-label use means that a medication is being used to treat a condition for which it was not approved. This could be related to the clinical picture of infected patients (electrolyte imbalance, proinflammatory state, disseminated intravascular coagulation status), high and prolonged doses of these drugs, polypharmacy and drug interactions, which would increase the risk of adverse events, which in some cases could be avoidable ⁵.

The implementation of pharmacovigilance in the context of the SARS-CoV-2 pandemic was essential, due to the high inpatient and outpatient demand for pharmaceutical products with limited scientific evidence on their benefits and safety in the treatment of COVID-19 ⁶. According to the World Health Organization (WHO), pharmacovigilance is the science and activities related to the detection, evaluation, understanding and prevention of adverse drug events, also known as adverse reactions, or any other drug-related problems ⁷.

Worldwide, several pharmacovigilance studies during the SARS-CoV-2 pandemic reported adverse events, such as prolongation of the QT interval and arrhythmias due to the use of HCQ^8,9, in addition to the synergic arrhythmogenic effects of the AZI/HQ combination ^10-14. Likewise, neurological ¹⁵ and gastrointestinal ¹⁶ adverse events have been reported for ivermectin, and serious hepatic effects ¹⁷ and neutropenia, superinfections, reactivation of latent infections, hepatitis and cardiac abnormalities for TOB ¹⁸.

In Peru, the Ministry of Health (MINSA), through the technical document “Prevention, Diagnosis and Treatment of persons affected by COVID-19”, recommended that the prescription of some of the aforementioned drugs in mild, moderate and severe cases of COVID-19, emphasizing that possible adverse reactions to these drugs should be monitored, since the safety information available on this disease is limited ¹⁹. However, little has been done in this regard, therefore, information on adverse events to these products is scarce in Peru.

Thus, the safety profile of “off-label” medications for COVID-19 in the Peruvian population needs to be described. Therefore, this study aims to characterize all reports of suspected adverse reactions associated with HCQ, AZI, TOB and IVM prescribed for the treatment of patients hospitalized for COVID-19 at the Edgardo Rebagliati Martins National Hospital (HNERM) and admitted to the Institutional Reference Center for Pharmacovigilance and Technovigilance of EsSalud (CRI-EsSalud), from April to October 2020.

MATERIALS AND METHODS

RESULTS

Between April to October 2020, we identified 154 notifications of COVID-19 hospitalized patients who had at least one AE during “off-label” pharmacological treatment of COVID-19 prescribed during their hospital stay at HNERM. A total of 183 AEs were identified, because 18.8% of patients had more than one AE. The median age of the affected patients was 59 (IQR: 50-70) years and 70.8% were male (109 out of 154). We found that most of the affected patients had no history of ADR (Table 1).

Most cases were moderate (53%; 82 out of 154), however, 10.4% were severe. It is important to note that 38 people who had an AE died. It was not possible to determine to what extent the drugs contributed to death (Table 1).

AE according to organ-system classification

According to the SOC classification of AE, 39.5% (n=72) were cardiovascular, 35.5% (n=65) were hepatobiliary, 16.9% (n=26) corresponded to gastrointestinal events. The remaining 8.1% (n=17) belonged to other SOC classes (Figure 1). Table 2 details the characteristics by AE.

Table 2 Characteristics of the reported and analyzed adverse events. 

Characteristics		Cardiovascular AE		Hepatobiliary AE		Gastrointestinal AE		Post-infusion AE		Hematological AE		Psychiatric Neurological AE
		n=72	%	n=65	%	n=26	%	n=9	%	n=4	%	n=7	%
Sex
	Female	19	26.6	19	28.6	10	38.7	-	-	1	25.0	2	28.6
	Male	53	73.4	46	71.4	16	61.3	9	100.0	3	75.0	5	71.4
Age (years)
	≤50	26	36.7	19	28.6	2	6.5	4	44.4	1	25.0	-	-
	>50	46	63.3	46	71.4	24	93.5	5	55.6	3	75.0	7	100.0
Medical history a
	HT	27	38.0	12	18.6	6	22.6	1	11.1	2	50.0	4	57.1
	T2DM	12	16.5	8	12.9	3	9.7	3	33.3	1	25.0	2	28.6
	Respiratory (TBC)	2	2.5	7	11.4	-	-	2	22.2	1	25.0	-	-
	Another	20	27.9	12	18.6	6	22.6	1	11.1	2	50.0	2	28.6
	None	15	20.3	33	50.0	13	48.4	2	22.2	0	0	1	14.3
Severity
	Mild	20	27.9	31	47.1	3	12.9	-	-	1	25.0	2	28.6
	Moderate	34	46.8	32	48.6	22	83.9	8	88.9	2	50.0	5	71.4
	Severe	18	25.3	2	4.3	1	3.2	1	11.1	1	25.0	-	-
Outcome
	Recovered	49	68.0	56	85.7	25	96.7	8	88.9	3	75.0	5	71.4
	Died	23	32.0	9	14.3	1	3.3	1	11.1	1	25.0	2	28.6
Suspected drugs
	AZI	17	24.1	13	20.0	19	77.4	-	-	1	25.0	5	71.4
	HCQ	12	16.5	10	15.7	1	3.2	-	-	-		-
	AZI + HCQ	36	50.6	7	11.4	-	-	-	-	-		-
	IVM (alone and combined)	-	-	9	14.3	2	6.5	-	-	-		2	28.6
	TOB	2	2.5	15	22.9	1	3.2	9	100.0	2	50.0	-	-
	Another b	5	6.3	10	15.7	3	9.7	-	-	1	25.0	-	-
AE causality
	Defined (ADR)	4	5.1	1	1.5	-	-	-	-	-	-	-
	Probable	42	58.2	24	37.1	8	29.0	3	33.3	2	50.0	2	28.6
	Possible	19	26.6	26	40.0	15	58.1	6	66.7	1	25.0	5	71.4
	Conditional	7	10.1	14	21.4	3	12.9	-	-	1	25.0	-	-

AE: adverse event; HT: hypertension; T2DM: type 2 diabetes mellitus; TB: tuberculosis; AZI: azithromycin; HCQ: hydroxychloroquine; IVM: ivermectin; TOB: tocilizumab; ADR: adverse drug reaction.

^a Patients may have more than one condition, ^b includes ranitidine, omeprazole, losartan, rendesivir, enoxaparin, ceftriaxone, linezolid, cefepime (possible interactions with AZT and HCQ).

Of the 72 reported cardiovascular AEs, 28 were cases of QT interval prolongation, of which 10 described QT interval ≥500 ms. Likewise, 37 cases of arrhythmias were reported (Figure 1). We found that 73.4% of the cases happened in men and 36.7% of the patients were 50 years of age or younger. In addition, 18 cases were severe and 23 patients died, but whether their death was causally related to AE could not be determined. We also found 36 reports of cardiovascular AEs describing the AZI/HCQ combination as a probable cause, 17 cases for AZI and 12 for HCQ. According to the causality assessment, 4 cases were confirmed and most were considered probable (58.2%).

Figure 1 Adverse events by the lowest level terms according to system-organ classification and severity. 

Sixty-five hepatobiliary AEs (laboratory alterations) were identified, 64 cases had transaminitis (51 cases with elevated AST and ALT, 11 with elevated ALT and two with elevated AST) and one with hyperbilirubinemia. We found that 71.4% were men and 28.6% of the patients were 50 years old or younger. Two cases were severe and nine patients died. TOB and AZI were the most suspected to produce hepatobiliary AEs. In 45.7% of the cases, upon discontinuation of the suspected drugs, laboratory tests returned to baseline levels. Most cases were considered possible (40%) and probable (37.1%), however, one case was confirmed.

Gastrointestinal AEs were reported in 26 cases (15 cases of diarrhea and 11 of nausea, vomiting and abdominal pain). One case of diarrhea was severe because it extended the patient’s hospital stay. The notifiers indicated that AZI was possibly related to these AEs. Eighty percent of the cases with diarrhea subsided after discontinuation of AZI. Nine patients had AEs after intravenous administration of TOB. Of the total cases, one patient with severe COVID-19 died within six days of post-infusion reaction.

Four case reports showed hematological involvement (intracranial hemorrhage, anemia and hematological laboratory alterations). It is important to note that 3 cases showed potential drug interactions (TOB/linezolid, extended dose enoxaparin/fluoxetine, extended dose enoxaparin/methylprednisolone). One case was severe and fatal. Most cases were related to TOB and were considered probable.

Three cases of psychiatric symptoms (psychomotor agitation, delirium and hallucinations) and four neurological AEs were reported (two cases of headache, one of dizziness and one of insomnia). All occurred in patients older than 50 years of age. Most were moderate. AZI was suspected as the main cause of 71.4% of AEs. Reports of headache and insomnia describe MVI as the cause.

DISCUSSION

This study characterized 154 reports of AEs associated with HCQ, AZI, TOB and IVM prescribed as “off-label” in the treatment of patients hospitalized for COVID-19 at HNERM. Most of the reports described male patients affected by at least one AE, this evidence is consistent with that described by other authors for the use of drugs without approved indication as treatment against COVID-19 ^8-10,21,22. A possible explanation is that, worldwide, men had more severe COVID-19, and therefore, more risk of being hospitalized ²³, of being exposed to pharmacological treatments against COVID-19 and consequently, they were more likely to have adverse reactions. It is important to mention that AEs to AZI were more frequent than to HCQ. However, most pharmacovigilance studies point to HCQ as the leading cause. This is possibly due to the fact that there was more AZI consumption than HCQ in EsSalud healthcare centers that reported at least one AE.

We identified 183 AEs from the 154 reports. Most of these were cardiovascular, with QT interval prolongation and arrhythmias standing out; the AZI/HCQ combination were the main cause of these AEs. One of the reasons for this statement is that the AZI/HCQ pharmacological interaction and its arrhythmogenic effects have been widely described during the COVID-19 pandemic ^8-14,24. The arrhythmogenic mechanism of both drugs is explained by the blockade of cardiac voltage-dependent potassium channels, which would prolong cardiac repolarization, leading to ventricular arrhythmia ²⁵. Additionally, AEs may occur due to drug-disease interaction, thus, the AZI/HCQ combination may prolong the QT interval, but this AE may be exacerbated by electrolyte imbalances in patients with COVID-19. Several drug regulatory agencies such as the European Medicines Agency (EMA) ²⁶, the Spanish Agency for Medicines and Health Products (AEMPS) ^27,28 and DIGEMID ²⁹ warned of the risks of arrhythmias due to AZI/HCQ consumption, suggesting intensive cardiological monitoring if these drugs were prescribed.

Hepatobiliary (transaminitis) and gastrointestinal AEs were identified in this study. According to the causality assessment, most cases were possibly associated to TOB y AZI (score of 5-6 according to the causality algorithm by Karch and Lasagna), because COVID-19 was considered an alternative cause to transaminitis and diarrhea. Thus, previous studies describe that those two symptoms are the most frequent in patients with COVID-19 and hepatobiliary and gastrointestinal disorders, especially during stage 2 of the disease ^30,31. However, it is important to note that the TOB medication guide describes both symptoms as frequent (≥1/100 to <1/10) ³² and that pharmacovigilance studies during the COVID-19 pandemic report hepatic AEs associated with TOB ^17,18. For AZI, diarrhea is described as very frequent (≥1/10) and transaminitis as rare (≥1/1000 to <1/100) ³³. Therefore, the association between this type of AEs and these pharmaceuticals, or with COVID-19 itself, cannot be ruled out.

The use of “off-label” medications can lead to errors during treatment which, in turn, could lead to the occurrence of AEs. In this study, we found AEs possibly associated with errors during treatment. Three cardiac AEs (QT interval prolongation, bradycardia and atrial fibrillation) were related to high doses and extended use of AZI (500 mg for seven days). We found one case of a subdural hematoma in a patient who received fluoxetine, enoxaparin (therapeutic dose for more than 21 days) and AZI 250 mg for 15 days. This drug interaction is considered dangerous and should be avoided ³⁴. These cases had fatal outcomes. Additionally, we found a case of moderate thrombocytopenia in a patient receiving linezolid and enoxaparin. It should be noted that this AE is to be expected with both drugs. Initially, patients with COVID-19 were exposed to several drug combinations, which makes it difficult to identify the drug or drug association responsible for the AE. Therefore, the pharmacological characteristics of drugs should be considered before prescribing them. Pharmacovigilance should also be implemented.

The high hospital demand for pharmaceutical products with limited or no scientific evidence for the treatment of COVID-19 was an opportunity for expanding pharmacovigilance in Peru; however, it was not properly implemented ³⁵. Both MINSA and the Instituto de Evaluación de Tecnologías en Salud e Investigación (IETSI) recommended pharmacovigilance for off-label treatment of COVID-19 ³⁶. We found 154 AE reports at the HNERM, with a reporting rate of 8%. Although this is an unusual number of AE reports in seven months for the studied drugs, this may be due to substantial underreporting. Underreporting of suspected ADRs is a worldwide problem. The lack of awareness of the impact of reporting, the demanding workload of health professionals, especially at the beginning of the pandemic, are possible reasons for underreporting. Another important cause for underreporting is that health professionals consider pharmacovigilance as an administrative function rather than a healthcare activity. Pharmacovigilance contributes to patient safety; therefore, it must be considered as a clinical and healthcare activity by the professionals responsible for the direct care of patients with COVID-19.

The retrospective design was one of the main limitations of our study. Although we had access to electronic pharmacovigilance medical records, in some cases the information was incomplete, possibly due to the high patient demand. We did not access the physical medical records or nursing charts, the latter being important because it contains information regarding the doses of the drugs administered to patients. Likewise, although the modified Karch and Lasagna algorithm is supported by the Peruvian Pharmacovigilance System, the evaluation can be subjective when assessing parameters such as the existence of alternative causes, contributing factors and complementary evaluations. Furthermore, this algorithm only allows us to see the degree of association between the adverse event and the medication, but does not define causality, therefore we cannot assume that the drugs administered during hospitalization caused the adverse events, even more so when the information regarding the medication received outside the hospital was not available. Even with these limitations, this study did not only describe the HNERM reports from the EsSalud pharmacovigilance database, but also collected clinical information to make a causality assessment and describe the AE outcomes. In addition, this is one of the few studies that describes AE cases due to the use of “off-label” medication against COVID-19 in Latin America.

In conclusion, we found AEs associated with the widespread use of drugs against COVID-19 without scientific evidence during the beginning of the pandemic, with cardiovascular and hepatobiliary conditions standing out. AZI and the AZI/HCQ combination were the ones that were reported to be associated with most safety issues. However, AEs were also related to TOB and IVM. Most of the cases were moderate and the patients recovered. Additionally, AEs were found to be underreported, therefore, health professionals treating patients with COVID-19 should recognize the importance of pharmacovigilance of “off-label” drugs, mainly TOB, and thus contribute to patient safety.