The World Health Organization (WHO) defines an adverse drug reaction (ADR) as any response to a drug that is noxious and unintended, and that occurs at doses normally used in man for the prophylaxis, diagnosis or therapy of disease, or for the modification of physiological function.[1] Because this definition excludes therapeutic failures, intentional overdose and medication error, it captures a category of harm that is intrinsic to appropriate prescribing rather than incidental to it. ADRs are therefore an unavoidable accompaniment of pharmacotherapy, and their systematic detection is one of the core activities of pharmacovigilance.[2]

The clinical and economic burden imposed by ADRs is substantial. A meta-analysis of prospective studies from the United States estimated that serious ADRs occurred in 6.7% of hospitalised patients and ranked ADRs between the fourth and sixth leading cause of death.[3] In a large prospective analysis of 18,820 admissions to two hospitals in Merseyside, Pirmohamed and colleagues found that 6.5% of admissions were related to an ADR, that the reactions were directly responsible for death in 0.15% of all patients admitted, and that the majority of these reactions were avoidable.[4] Reactions arising during a hospital stay prolong the duration of admission, add investigation and treatment costs, erode confidence in the treating team and, in a publicly funded setting, divert scarce resources away from other patients.[5]

Surgical wards present a distinctive risk environment. Patients are exposed within a short admission to several pharmacological classes simultaneously — perioperative and therapeutic antimicrobials, opioid and non-opioid analgesics, intravenous crystalloids, proton pump inhibitors, antiemetics and anaesthetic agents — so that the opportunity for both dose-related toxicity and drug interaction is high. Many admissions are emergencies in which a detailed drug and allergy history cannot be obtained, and a substantial proportion of patients carry comorbidities such as diabetes mellitus that alter drug handling and wound healing. Studies conducted specifically in surgical in-patients have consistently reported that antimicrobials and analgesics dominate the causative drug profile and that gastrointestinal reactions predominate.[6,7]

Despite this, ADRs in India remain considerably under-reported. Spontaneous reporting, the mainstay of the Pharmacovigilance Programme of India, depends on the willingness of the treating clinician, and reporting rates from surgical disciplines are lower than those from internal medicine and paediatrics.[6] Reporting alone, however, is of limited value unless each report is subsequently characterised. Assessment of causality establishes the strength of the association between drug and event; assessment of predictability distinguishes dose-related, pharmacologically foreseeable Type A reactions from idiosyncratic Type B reactions; assessment of severity determines the clinical consequence; and assessment of preventability identifies the subset of reactions that could have been averted by a change in prescribing practice.[8–11] It is this last dimension that converts a passive surveillance exercise into an actionable quality-improvement intervention.

Data describing these four dimensions together, from a surgical in-patient population in a public tertiary care teaching hospital in India, are limited. The present study was therefore undertaken to analyse the pattern of spontaneously reported ADRs among in-patients of the Department of Surgery, and to characterise them with respect to causality, predictability, severity and preventability.

This was a cross-sectional study conducted in the Department of Surgery, Victoria Hospital, attached to Bangalore Medical College and Research Institute (BMC&RI), Bengaluru. The institution is a tertiary care teaching hospital and hosts an ADR Monitoring Centre (AMC) recognised under the Pharmacovigilance Programme of India. The study was carried out over a period of six months. Ethical considerations The study protocol was reviewed and approved by the Institutional Ethics Committee of BMC&RI before commencement. As the study analysed reports generated through routine spontaneous pharmacovigilance activity and involved no intervention or change in patient management, no additional investigation was performed on any patient. Patient identifiers were removed at the point of data entry and all records were handled confidentially. Source of data and participants The source of data was the set of suspected ADR reports received at the AMC from in-patients admitted under the Department of Surgery during the study period. All spontaneously reported suspected reactions occurring in surgical in-patients of either sex and any age were included. Reports were excluded if the description of the reaction was incomplete, if the suspected drug was not identifiable, or if the event was attributable to intentional overdose, deliberate self-harm, drug abuse, therapeutic failure or a documented medication error, since such events fall outside the WHO definition of an ADR. Data collection Reactions were reported by treating clinicians, nursing staff and postgraduate residents, and were supplemented by daily ward visits by the investigator to encourage and validate reporting. For every report, details were entered into the CDSCO suspected adverse drug reaction reporting form (version 1.2). Information recorded comprised patient demographics (age, sex), admitting surgical diagnosis and whether the patient was in the preoperative or postoperative phase, relevant comorbidities, a complete list of concomitant medications with dose, route and duration, a description of the reaction with dates of onset and of resolution, the suspected drug or drugs, action taken with respect to the suspected drug, treatment administered for the reaction and the eventual outcome. Dechallenge information was recorded where available; rechallenge was not performed for ethical reasons. Assessment of reactions Each documented reaction was assessed independently along four dimensions. Causality was determined using the WHO–Uppsala Monitoring Centre causality assessment system, which categorises reactions as certain, probable, possible, unlikely, conditional or unassessable.[8] Predictability was classified according to the Rawlins and Thompson system into predictable, dose-dependent Type A reactions and unpredictable, non-dose-dependent Type B reactions.[9] Severity was graded using the Modified Hartwig and Siegel scale as mild (levels 1–2), moderate (levels 3–4) or severe (levels 5–7).[10] Preventability was assessed using the Modified Schumock and Thornton criteria and classified as definitely preventable, probably preventable or not preventable.[11] Reactions were also grouped by the organ system affected and the suspected drugs classified pharmacologically. Assessments were performed by the investigator and independently verified by a faculty member of the Department of Pharmacology; disagreements were resolved by discussion until consensus was reached. Statistical analysis Data were entered into a Microsoft Excel spreadsheet and analysed descriptively. Categorical variables are expressed as frequencies and percentages, and continuous variables as mean ± standard deviation. Percentages were calculated using the total number of reported reactions (n = 60) as the denominator unless stated otherwise. No inferential statistical testing was undertaken, as the study was descriptive in intent.

A total of 60 suspected adverse drug reactions were reported from the Department of Surgery during the one-year study period and were available for analysis. The demographic profile of the patients is presented in Table 1. The mean age was 41.82 ± 15.67 years, and reactions were reported slightly more often in women (56.7%) than in men (43.3%), broadly mirroring the sex distribution of admissions to the unit.

Table 1. Demographic profile of patients experiencing adverse drug reactions (n = 60)

Table 2. Distribution of adverse drug reactions by clinical diagnosis and phase of care (n = 60)

Reactions were distributed almost equally between the preoperative and the postoperative phase of care, 31 (51.7%) and 29 (48.3%) respectively (Table 2). Among preoperative patients, diabetic foot ulcer was the commonest admitting diagnosis (16.7%), a group in whom prolonged broad-spectrum antimicrobial therapy is routine. Among postoperative patients, hernioplasty (16.7%) and appendicectomy (13.3%) together accounted for the largest share, reflecting the case-load of a general surgical unit.

The pattern of individual reactions is shown in Table 3. Diarrhoea was the single most frequent reaction, accounting for one-fifth of all reports, followed by vomiting and headache (15.0% each). Dysgeusia, reported in 10.0% of cases, was associated exclusively with metronidazole. Cutaneous reactions were confined to rash with itching, and one metabolic reaction (hyponatraemia) was recorded.

Table 3. Pattern of adverse drug reactions reported (n = 60)

Grouped by organ system (Table 4), the gastrointestinal tract was involved in two-thirds of all reactions, the central nervous system in just under a quarter, the skin in 8.3% and electrolyte balance in a single case. This distribution follows directly from the causative drug profile described below.

Table 4. System-wise distribution of adverse drug reactions (n = 60)

Eleven drugs or drug combinations were implicated (Table 5). Ceftriaxone was the commonest single suspect (23.3%), followed by metronidazole and tramadol (20.0% each). When grouped pharmacologically (Table 6), antimicrobials accounted for 65.0% of all reactions and analgesics for a further 25.0%, so that these two classes together explained nine out of every ten reactions reported.

The results of the four assessments are summarised in Table 7. Causality was categorised as ‘possible’ in 85.0% and ‘probable’ in 15.0% of reactions; no reaction met the criteria for ‘certain’, since rechallenge was not undertaken. The great majority of reactions (93.3%) were predictable, dose-related Type A reactions. Severity assessment classified 86.7% as mild and 13.3% as moderate, with no severe reaction recorded. Preventability assessment identified 76.7% of reactions as definitely preventable and a further 3.3% as probably preventable, so that four out of five reactions were considered avoidable.

Table 5. Individual drugs implicated in the reported adverse drug reactions (n = 60)

Table 6. Class of drugs implicated in the reported adverse drug reactions (n = 60)

Table 7. Causality, predictability, severity and preventability of the reported adverse drug reactions (n = 60)

Management and outcome are presented in Table 8. No specific treatment was required for 60.0% of reactions; 30.0% were managed by non-pharmacological measures such as reassurance, dietary modification, hydration or slowing the rate of infusion, and 10.0% required additional drug therapy, so that 40.0% of reactions in total required some form of active management. The suspected drug was withdrawn in 15.0% of cases and continued, with or without dose modification, in the remaining 85.0%. At the time of reporting, 61.7% of patients had recovered completely and 38.3% were still recovering; there was no instance of a reaction leaving a residual deficit, and no fatal reaction occurred.

Table 8. Management of the adverse drug reaction and patient outcome (n = 60)

In this one-year analysis of spontaneously reported reactions from the surgical wards of a public tertiary care hospital, 60 ADRs were characterised. The mean age of 41.82 ± 15.67 years and the modest female preponderance (56.7%) are consistent with the profile reported by Yadav and Acharya from surgical in-patients, and reflect the age distribution of the general surgical case-mix rather than any age-specific susceptibility.[7] A female preponderance has been described repeatedly in Indian ADR studies and has been attributed to differences in body composition, pharmacokinetics and health-seeking behaviour rather than to differential exposure.[12] The gastrointestinal tract was the organ system most frequently involved (66.7%), and diarrhoea was the single commonest reaction (20.0%). This distribution is a direct consequence of the causative drug profile: antimicrobials accounted for 65.0% of reactions, with ceftriaxone (23.3%) and metronidazole (20.0%) the leading individual agents. Broad-spectrum cephalosporins disturb the colonic microbiota and are a well-recognised cause of antibiotic-associated diarrhoea, while metronidazole characteristically produces dysgeusia, nausea and a metallic taste — dysgeusia was reported in 10.0% of our cases. Analgesics, principally tramadol, accounted for a further 25.0% of reactions and explain the cluster of central nervous system events (headache, dizziness) that constituted 23.3% of the total. A comparable predominance of antimicrobials and analgesics in surgical wards was reported by Sowmyanarayan and Banerjee and by Yadav and Acharya.[6,7] Jose and Rao, analysing spontaneous reports across an Indian teaching hospital, likewise found antimicrobials to be the most frequently implicated class.[13] Causality was ‘possible’ in 85.0% of reactions and ‘probable’ in the remainder, with no reaction categorised as ‘certain’. This is an expected finding in a spontaneous reporting programme: the ‘certain’ category of the WHO–UMC system requires a satisfactory rechallenge, which cannot ethically be performed, and the high prevalence of polypharmacy in surgical patients makes it difficult to exclude a competing explanation for any given event.[8] The preponderance of ‘possible’ assessments should therefore be read as a limitation inherent to the method rather than as evidence of weak association. The most actionable findings concern predictability, severity and preventability. Nearly all reactions (93.3%) were predictable Type A reactions, and 86.7% were mild, requiring at most withdrawal of the drug or symptomatic treatment; no severe or fatal reaction occurred. Critically, 76.7% were judged definitely preventable. Because Type A reactions are pharmacologically foreseeable and dose-related, this proportion represents harm that could plausibly be reduced through narrower empirical antimicrobial choice, adherence to institutional surgical prophylaxis protocols that limit the duration of perioperative cover, dose adjustment for renal function, and routine documentation of drug allergy at admission.[4,11] Comparable preventability estimates have been reported from other Indian centres.[12,13] The study has limitations. Spontaneous reporting captures only a fraction of reactions occurring in a ward and is biased towards events that are early in onset, clinically obvious and already associated with the suspect drug, so the true incidence is certainly higher and rare Type B reactions are under-represented. The absence of a denominator of total drug exposures precludes calculation of incidence rates, and the single-centre design and modest sample limit generalisability. Prospective intensive monitoring in a defined cohort would address several of these constraints.

Adverse drug reactions in the surgical wards of this tertiary care hospital were predominantly mild, predictable and preventable Type A reactions, affecting chiefly the gastrointestinal tract and attributable in almost two-thirds of instances to antimicrobials, with analgesics accounting for most of the remainder. No severe or fatal reaction was encountered, and every patient recovered or was recovering at the time of reporting. The finding that more than three-quarters of reactions were definitely preventable is the central message of this study: it indicates that a substantial share of the drug-related harm observed in surgical practice is amenable to correction through rational prescribing, restraint in the choice and duration of empirical antimicrobial therapy, and careful attention to drug history and dose individualisation. Sustained ADR monitoring creates awareness among clinicians, nurses, pharmacists and patients, generates institution-specific safety data to inform local prescribing policy, and thereby improves the quality of health care delivered and ensures patient safety.

Declarations

Funding: No funding sources.

Conflict of interest: None declared.

Ethical approval: The study was approved by the Institutional Ethics Committee, Bangalore Medical College and Research Institute, Bengaluru.

World Health Organization. International drug monitoring: the role of national centres. Report of a WHO meeting. World Health Organ Tech Rep Ser. 1972;498:1-25.

2. World Health Organization. The importance of pharmacovigilance: safety monitoring of medicinal products. Geneva: World Health Organization; 2002.
3. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA. 1998;279(15):1200-5.
4. Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ, et al. Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ. 2004;329(7456):15-9.
5. Ramesh M, Pandit J, Parthasarathi G. Adverse drug reactions in a south Indian hospital — their severity and cost involved. Pharmacoepidemiol Drug Saf. 2003;12(8):687-92.
6. Sowmyanarayan S, Banerjee S. Monitoring of adverse drug reactions in medicine, paediatric and surgical departments of a tertiary care hospital: a prospective observational study. Int J Basic Clin Pharmacol. 2018;7(4):778-82.
7. Yadav D, Acharya RP. Incidence and severity associated with adverse drug reactions in surgery inpatients. J Pharm Sci Res. 2015;7(9):671-5.
8. Uppsala Monitoring Centre. The use of the WHO-UMC system for standardised case causality assessment. Uppsala: Uppsala Monitoring Centre; 2018.
9. Rawlins MD, Thompson JW. Mechanisms of adverse drug reactions. In: Davies DM, editor. Textbook of adverse drug reactions. 4th ed. Oxford: Oxford University Press; 1991. p. 18-45.
10. Hartwig SC, Siegel J, Schneider PJ. Preventability and severity assessment in reporting adverse drug reactions. Am J Hosp Pharm. 1992;49(9):2229-32.
11. Schumock GT, Thornton JP. Focusing on the preventability of adverse drug reactions. Hosp Pharm. 1992;27(6):538.
12. Jose J, Rao PG. Pattern of adverse drug reactions notified by spontaneous reporting in an Indian tertiary care teaching hospital. Pharmacol Res. 2006;54(3):226-33.
13. Edwards IR, Aronson JK. Adverse drug reactions: definitions, diagnosis, and management. Lancet. 2000;356(9237):1255-9.