Vaccine safety surveillance informs public health policy beyond immunization: A case-series on bleeding following vaccination, Nepal, 2016-2018.

BACKGROUND: Surveillance for adverse events following immunization (AEFI) is important to monitor vaccine safety and should lead to appropriate responses to improve health and immunization program. Bleeding following vaccination is not recognized as an important AEFI. Without policy of vitamin K (VK) prophylaxis at birth, vitamin K deficiency bleeding (VKDB) could be an important cause of bleeding in young infants and may manifest as AEFI.
METHODS: We retrospectively analysed all serious AEFI cases that presented with external or internal bleeding reported to Nepal's AEFI surveillance system during 2016-2018. The cases were classified as VKDB, suspected VKDB or non-VKDB.
RESULTS: During the period, 16 serious AEFI with symptom or sign of bleeding were reported representing 21.3% of all serious AEFI reported. Cases were between 40 and 94 days of age. The National AEFI Investigation Committee classified all cases as coincidental. Fourteen cases (87.5%) had bleeding from injection site. Median time from vaccination to injection site bleeding was 4.3 h (interquartile range: 2.1-11.6 h). Six cases (37.5%) had intra-cranial haemorrhage. Only one case had confirmed history of receiving VK at birth. Ten cases (62.5%) received appropriate treatment (VK injection; blood transfusion if needed). Based on limited laboratory investigations available, three cases (18.75%) could be classified as late onset VKDB and 11 cases (68.75%) as suspected late onset VKDB.
CONCLUSION: VKDB should be suspected in young infants presenting with bleeding including following vaccination, and prompt treatment should be initiated. Bleeding following vaccination should be recognized as an important AEFI as even a small amount of blood loss in young infants can be catastrophic. We posit that this series is a small subset of VKDB cases in Nepal detected through AEFI surveillance system. In countries without policy of VK prophylaxis at birth including Nepal, the policy should be introduced.

Introduction

The National Immunization Programme (NIP) of Nepal under Ministry of Health and Population (MoHP) has consistently achieved high coverage with the scheduled basic vaccines for the infant population [1]. The most recent nationwide demographic and health survey found that only 1% of the children were completely left out of reach of the NIP [2].

Nepal has an established surveillance system for reporting and investigating adverse events following immunization (AEFI). All reported serious AEFI cases are investigated, classified and feedback given as per the national and global AEFI guidelines [3], [4]. Causality assessment is conducted by the National AEFI Investigation Committee (AEFI-IC) using WHO recommended methods for causality assessment of an AEFI [5]. The AEFI-IC is now mandated by Nepal’s Immunization Act 2072 (CE 2016).

The Immunization Preventable Diseases unit (WHO-IPD) of World Health Organization, Nepal, with its nationwide network of surveillance medical officers (SMO) and a central coordinating office provides technical assistance to the process of serious AEFI case investigation, data compilation, analysis and presentation to the National AEFI-IC. A thorough AEFI case investigation is followed by review, causality assessment and classification of cases following standard WHO guidelines [4], [5].

Between 2016 and 2018, the AEFI surveillance system in Nepal identified a series of cases of bleeding in young infants presenting as adverse event following immunization. The National AEFI-IC classified all of these cases as coincidental events occurring after a vaccine injection, with a primary or suspected diagnosis for the majority of cases as late onset vitamin K deficiency bleeding (VKDB).

One of the preventable causes of infant mortality is VKDB (previously known as haemorrhagic disease of the newborn), a coagulopathy due to insufficient vitamin K stores in infants. VKDB is classified in the literature as early, classic, and late depending on age at onset of manifestations. Early VKDB presents within 24 h after birth, classic VKDB presents between 1 and 7 days after birth, and late VKDB presents in 2–12 (but can extend up to 24) weeks of age [6], [7]. Cases of late onset VKDB often present with intracranial haemorrhage with potential fatalities [8]. The preventative intervention for VKDB is administration of vitamin K shortly after birth and is a WHO recommended essential newborn care approach [9]. Administration of vitamin K at birth has been recommended since 1961 by the American Academy of Pediatrics [10]. However, in many low-income countries this recommendation may not have been adopted or implemented only in the private sector.

The incidence of classical VKDB in the absence of vitamin K prophylaxis ranges from 10 to 1500 per 100,000 neonates, and that of late VKDB ranges from 5 to 80 per 100,000 [11]. The exact magnitude of VKDB in Nepal is not known as there is no nationwide system of identifying, investigating and reporting VKDB. Currently there is no national policy in Nepal to administer vitamin K at birth. Nepal’s Every Newborn Action Plan, MoHP, does not include vitamin K administration at birth [12].

In this report we demonstrate how a sensitive nationwide vaccine safety surveillance system identified a potentially preventable but underreported cause of infant mortality. This information and recommendation made subsequently should help Ministry of Health and Population, Nepal, in making policy decision of introducing a non-vaccine intervention (vitamin K prophylaxis at birth) to lower infant mortality rate in Nepal from its current level of 32 per 1000 live births [2].

Methods

The NIP of Nepal provides BCG vaccine at birth; DTwP-HepB-Hib and OPV at 6, 10, and 14 weeks; fractional dose of IPV (fIPV) at 6 and 14 weeks (intramuscular IPV given at 14 weeks available until third quarter of 2016; fIPV started from third quarter of 2018); PCV at 6, 10 weeks, and 9 months; MR at 9 months and 15 months; JEV at 12 months; and Td during pregnancy. In July 2020, rotavirus vaccine was introduced in the NIP, which is given at 6 and 10 weeks of age. For vaccination as per NIP schedule for children, there are a total of seven scheduled contacts for seven series of vaccination with each contact having one or more injectable vaccines. An infant’s first immunization contact with an injectable vaccine is either at birth (with BCG vaccine) or at six weeks of age (with BCG, DTwP-HepB-Hib, fIPV, and PCV), followed by contacts at 10 and 14 weeks before the child reaches six months of age. Serious AEFI cases are reported, investigated and causality assessment is conducted based on national and WHO guidelines [3], [4], [5].

We retrospectively analysed all serious AEFI cases reported to the system that presented with external or internal bleeding in the past three calendar years (1 January 2016 – 31 December 2018) and were classified by the AEFI-IC. AEFI investigation reports and causality assessment documents of all such cases were accessed and reviewed. The anonymized cases were analysed for demographics, AEFI event details, health seeking behaviour, laboratory investigations conducted, treatment provided, diagnosis and outcome. No further data collection was undertaken. Analysis was conducted only on data that was available from routine AEFI case investigation and causality assessment done by AEFI-IC.

AEFI case investigation and causality assessment are part of standard public health surveillance under the National Immunization Program of Nepal. Therefore, individual consent from parent or caregiver is not deemed necessary for an AEFI case investigation.

Our primary aim was to investigate whether these cases of bleeding were vitamin K deficiency bleeding. Therefore, for purposes of classifying these bleeding cases by aetiology, we applied the following criteria:

Vitamin K deficiency bleeding

Diagnosis of VKDB was given if the case had prothrombin time (PT) greater than or equal to four times the normal value or International Normalized Ratio (INR) ≥ 4 with normal platelet count for age, and correction of the PT to normal levels after administration of vitamin K. These criteria [13], [14] were used to confirm diagnosis of VKDB. In these criteria, fibrinogen level test was not included as none of the case had fibrinogen test done.

Suspected VKDB

Any case with strong clinical suspicion of VKDB (that is, probable clinical diagnosis of VKDB given by AEFI-IC during causality assessment process), but not fulfilling the criteria given above due to unavailability of required relevant tests, and without evidence of other cause of bleeding, was given diagnosis of suspected VKDB.

Non-VKDB (other bleeding cases)

All remaining cases not fitting criteria given above for ‘VKDB’ or ‘suspected VKDB’ and having evidence of other probable cause of bleeding were classified as non-VKDB cases.

We used median and interquartile ranges for statistical measures. Microsoft ® Excel ® 2016 was used for data calculations, and graphic outputs.

Results

Between 1 January 2016 and 31 December 2018, a total of 75 serious AEFI cases were reported to the surveillance system, all of which were investigated and causality assessment was conducted. This represents an average annual rate of 3.9 serious AEFI cases per 100,000 surviving infants. During this period, 15,350,165 injectable vaccine doses, and 4,865,954 oral vaccine doses were given through the National Immunization Program as reported by the Health Management Information System. The reported serious AEFI cases represent an average annual rate of 0.37 serious AEFI cases per 100,000 vaccine doses.

Among total serious AEFI cases, six cases (17.6%) in 2016, four cases (26.7%) in 2017, and six cases (23%) in 2018, presented with bleeding as main symptom or sign. In total, there were 16 serious AEFI cases (21.3%) with bleeding symptom or sign.

Background characteristics and onset of symptoms

The cases of bleeding were reported from 13 out of 75 districts. There was no clustering of cases. Ten districts reported one case each. Three districts reported two cases each; however, each of the two cases in each of these three districts was reported in different years. All these 16 cases were classified as coincidental events not causally related to the putative vaccine doses(s) by National AEFI-IC.

Among these total 16 cases, 11 (69%) were male. All cases were between 40 and 94 days of age, with a median age of 57 days and an inter quartile range (IQR) of 44–66 days (Fig. 1) at onset of AEFI. Eleven cases (69%) had institutional delivery, three (19%) had home delivery, and delivery place was unknown for two cases (12%). Only one case (6%) had confirmed history of receiving vitamin K at birth, whereas nine cases (56%) had confirmed history of not receiving vitamin K at birth (Table S1). For the remaining six cases (38%), receipt of vitamin K at birth was unknown.

Fig. 1

Age at onset of AEFI cases presenting with symptom or sign of bleeding, Nepal, 2016–2018.

Out of the total cases, 14 (87.5%) cases had bleeding from injection site, among which 13 cases had bleeding from injection site as the first symptom, whereas the remaining one case had blood in stool as the first symptom. The two remaining cases without evidence of bleeding externally, one had fever as first symptom and was clinically suspected to have intracranial haemorrhage (ICH) because of neurological signs and raised PT (INR 2.1), whereas the other had swelling of one thigh and excessive cry as first symptoms and had evidence (CT scan) of ICH (Table S1).

Among the 14 cases with evidence of bleeding externally, the majority (10 cases; 71.4%) had bleeding symptoms after receiving second series vaccinations (all or part of DTwP-HepB-Hib, PCV, fIPV, OPV first doses), two cases had bleeding symptoms after receiving combined first (BCG) and second series vaccinations together, and two cases had bleeding symptoms after receiving third series vaccinations (DTwP-HepB-Hib, PCV, OPV second doses). The remaining two cases without evidence of external bleeding had the second series (DTwP-HepB-Hib, PCV, OPV first doses) and the fourth series (DTwP-HepB-Hib, OPV third doses) vaccinations respectively. All cases occurred after at least one intramuscular vaccine injection.

Of the 14 cases with bleeding from injection site, the range of time of onset of injection site bleeding after vaccination was between 30 min and 24 h (one outlier case had onset at 48 h), with median time of 4.3 h (IQR: 2.1–11.6 h). Eight out of the 14 cases (57%) had onset of injection site bleeding within six hours and 11 (78.6%) had bleeding within 12 h of vaccination (Fig. 2). For the two cases without injection site bleeding but with suspected and confirmed ICH, the first symptom was fever (72 h post vaccination) and swelling in left thigh with fever (one hour post vaccination) respectively.

Fig. 2

Time of onset of injection site bleeding after vaccination, serious AEFI cases, Nepal, 2016–2018.

Health-care seeking behaviour, hospitalization and care received at the health facility

The first contact for care for nine cases (56.3%) was public health facility, out of which two were brought to district hospital, one to zonal hospital, and six to either health post or primary health centre. Six cases (37.5%) first contact was with private clinic, whereas the remaining one case (6.2%; the case without external bleeding and fever as first symptom, suspected ICH, and seizure) was taken to a traditional healer.

Among the 14 cases with evidence of external bleeding, four cases sought medical care or advice immediately or within one hour after noticing the symptom, depending on the actual or subjective severity of bleeding. However, for rest of the cases, the range for seeking medical care was between 6 h and 4.5 days. Only two cases sought medical care after 24 h of bleeding when infant became seriously ill (after 2.9 and 4.5 days respectively), the reason being that the parents thought the bleeding was mild. For the two cases without evidence of external bleeding, one case sought medical care immediately after noticing fever and swelling in thigh, whereas the other case sought proper medical care only after seven days of onset of fever later developing cough, fast breathing, and seizures. This child was taken to traditional healers before being taken to a hospital.

The minimum number of visits for health care or health facility visits was two and maximum was five. For seven cases (43.7%) the number of visits were two, for other seven cases (43.7%) the visits were three, and for one case each the visits were four and five respectively. The need for visiting multiple health facilities was necessitated by not receiving proper intervention or not having correct diagnosis in the initial visit(s) or due to referral to higher medical centre. Among the total, seven cases also had to stop at home in-between visits to health facilities. The reasons were that five cases were sent back home or discharged as injection site bleeding stopped and condition of the child seemed to improve, one case did not have external bleeding and was sent home from OPD with symptomatic treatment for thigh swelling and excessive cry, and one case was referred to higher centre but the parents took the child home hoping that symptoms will subside.

Fourteen cases were hospitalized (admitted). The median time to hospitalization from onset of symptoms was 25.5 h (range: 12 h − 6 days). The remaining two cases who were not hospitalized, one presented in OPD but was referred to higher centre (and later died), and the other expired after 15 min of arriving to a hospital. The median duration of admission in a hospital was three days (range: 0.3–17 days). Only 10 cases (62.5%) received appropriate treatment (vitamin K injection; blood transfusion if needed), and one case received blood transfusion only, and one other case received blood transfusion but the receipt of vitamin K was doubtful. None of these 10 cases received appropriate treatment on the first contact with a health facility. Five cases (50%) received appropriate treatment on second contact, four (40%) received on third contact, and one (10%) received on the fifth contact to a health facility.

Intracranial haemorrhage (ICH)

Six cases (37.5%) out of total 16 cases had evidence of ICH, found either through CT scan (four cases) or autopsy (two cases) (Table S1). The remaining 10 cases did not have CT scan or autopsy (if resulted in mortality) done; therefore, ICH could not be confirmed or excluded in these cases. Of the 14 hospitalized cases, four cases had CT scan done, all of which showed ICH. Of the nine total death cases, two had autopsy done, both of which showed ICH. Out of the six cases with evidence of ICH, five had injection site bleeding.

Laboratory tests

Eleven cases had blood haemoglobin measurement done, out of which seven had the test repeated after receiving treatment (Table S1). The range of haemoglobin among the eleven cases at presentation was 3.2–7.8 gm% (median: 6.2 gm%). All seven cases with repeat haemoglobin tests had received appropriate treatment (for one case, vitamin K receipt is doubtful), and the repeat test showed improvement in haemoglobin level except one case who received the treatment very late (after 7 days of onset) and did not survive.

Seven cases had PT tests done (activated partial thromboplastin time/APTT done in two cases) at presentation, among which five had repeat tests done after treatment (Table 1). All repeat tests showed improvement compared to initial tests. Among the total cases, 10 had platelet count test done before treatment. The lowest platelet count was 170,000/μL and the highest was 825,000/μL (Table S1).

Table 1

Blood coagulation tests, treatment and outcome, bleeding AEFI cases, Nepal, 2016–2018.

S. N	PT/APTT first test	Platelet count/μL	Treatment (VK and BT)	PT/APTT after treatment	Survived
1	Not done	323,000	Yes	PT 9 s, control 11 s, INR 0.83; APTT 25 s, control 28 s	Yes
2	PT 21 s, control 14 s	170,000	Yes (late)	Not done	No
3	PT 30 s, control 14 s, INR 2.1	811,000	Only VK (very late)	Not done	No
4	PT 18 s, control 13 s, INR 1.41	357,000	Only VK	PT 15 s, control 13 s, INR 1.16; APTT 38 s, control 26 s	Yes
5	PT 74 s, INR 6.24 s; APTT 138 s	470,000 (post treatment)	Yes (very late)	PT 18 s, control 14 s; APTT 35 s	No
6	PT 14 s	625,000	Yes	PT 11.2 s, control 12 s, INR 0.96; APTT 22.5 s	Yes
7	PT 1127.3 s, INR 10.07; APTT 1095.2 s	290,000	Yes (late)	PT 15 s, INR 1.31; APTT 32.35 s	Yes
8	PT 191.2 s, INR > 10 s	312,000	Yes (late)	PT 14.07 s, INR 1.14	Yes

S.N: serial number, PT: prothrombin time, APTT: activated partial thromboplastin time, VK: vitamin K, BT: blood transfusion, s: second.

Only cases with PT and/or APTT tests done are included in the table. Values as in the available reports are given in the table.

Treatment: within 2 days of onset; late: 3–5 days of onset; very late: after 7 days of onset.

Diagnosis (Classification)

Three cases (18.75%) fulfilled criteria to be classified as VKDB. All these three cases had pre-treatment PT > 4 times the normal value, received vitamin K, and showed rapid normalization of PT following vitamin K receipt (Table S1). Two cases had platelet count test done before treatment which was normal, and one case only had the test post-treatment which was normal. Out of these, two cases survived, and the remaining case that died had received appropriate treatment only after seven days of onset of symptoms.

Eleven cases (68.75%) fulfilled criteria to be classified as suspected VKDB. In these cases, there was no evidence of other aetiology of bleeding and had clinical suspicion of having VKDB. However, due to absence of relevant laboratory testing, these cases could not be classified as definitive VKDB. Altogether, 14 cases (87.5%) could be classified as late onset VKDB or suspected VKDB.

The remaining two cases (12.5%) were classified as non-VKDB. The diagnosis for these cases were ‘disseminated intravascular coagulation with suspected ICH’, and ‘suspected inborn error of metabolism’ respectively.

Outcome

In total, nine cases (56.3%) resulted in death (Table S1). Among the 14 cases with diagnosis of VKDB or suspected VKDB, eight cases (57%) died. The median time of death after onset of symptoms for the cases resulting to death was 68.5 h, and the range was 24 h to 17 days.

Discussion

Through this case-series, we have reported a series of bleeding cases after intramuscular vaccination presenting as AEFI identified through national surveillance system for vaccine safety. The majority of these cases were diagnosed or suspected (due to absence of relevant diagnostic testing) as late onset VKDB. All cases were classified after causality assessment by the AEFI-IC as coincidental cases (inconsistent causal association to immunization). Nevertheless, it demonstrates that the AEFI surveillance system in Nepal is capable of systematically picking up vaccine safety signals, investigating and classifying cases for vaccine safety as well as other serious health outcomes coincidentally related to vaccine administration.

Besides intracranial and gastrointestinal sites, skin is also well recognized as a common bleeding site for late onset VKDB [15]. Bleeding (prolonged or continuous bleeding at injection site and/or internal haemorrhage) following vaccination with injectable vaccines is not recognized as an important AEFI in literature. A case report of bleeding from injection sites of vaccines due to VK deficiency has been reported previously [16]. To our knowledge, our study is the first to report on multiple cases of bleeding presenting as adverse event following immunization. The bleeding cases during the three years period (2016–2018) represented significant proportion (1/5th) of all serious AEFI cases reported.

For this case-series, the minimum number of health facility visits was two and maximum was five. None of these cases received appropriate treatment (VK and if needed, blood transfusion) at first contact with a health facility, and only 62.5% actually received appropriate treatment. This is probably because of not suspecting diagnosis of VKDB, especially in the context of its manifestation around or post vaccination and beyond the neonatal period. At present there is no national policy in Nepal for prophylactic VK at birth possibly signifying that the disease is not yet recognized to have enough public health importance to merit a universal and affordable intervention.

Our study has several limitations. This is a retrospective analysis of events coincidentally discovered by the AEFI surveillance system. The number of cases found are limited within the expected cases and age group to be captured by the AEFI surveillance system. Therefore, our study does not represent the true incidence or burden of late VKDB in Nepal. It seemed there was strong time clustering of the onset of bleeding with 78.6% of the cases occurring within 12 h of an intramuscular injection. There was also a strong association between injection site bleeding and ICH (5 out of 6) where diagnostic testing for ICH or autopsy was undertaken. However, we could not explore statistical significance for these associations as the number of cases were relatively less.

As in most low-income settings, not all cases detected had a full diagnostic work-up to enable us to establish a definitive diagnosis of VKDB on biochemical parameters. Only seven cases had PT tests done at presentation, five of which had repeat tests after treatment. In our setting, PT/INR tests are not available in many health facilities. Fibrinogen level tests [13], [14], [17], measurement of specific vitamin K-dependent factors (II, VII, IX, X) [13], [17] or tests for proteins induced by the absence of VK [14] were not done in any of the case as they are largely unavailable. We could not obtain definitive history of VK administration at birth or lack thereof in 38% (6/16) of cases. However, as part of AEFI surveillance, available information was compiled as diligently as possible including autopsy reports wherever available. We have not conducted follow-up of the cases to find long-term sequelae including adverse neurodevelopment outcomes. Due to laboratory tests limitation, other causes of bleeding in young infants were not explored.

A recent systematic review [18] shows that the median (IQR) burden of late VKDB was 35 (10.5–80) per 100,000 live births in infants who had not received VK prophylaxis at birth. The burden was much higher in low and middle-income (LMIC) countries as compared with high-income countries [80 (72–80) vs 8.8 (5.8–17.8) per 100,000 live births]. Applying this higher median rate to the approximate Nepal birth cohort of 630,000, we project an incidence of 504 cases of late onset VKDB per year. Few children receiving VK at birth through private sector may not appreciably lower the burden. Studies have shown relatively higher incidence of late onset VKDB in some Asian countries [19], [20]. This could also be the case in Nepal. Late onset VKDB has a 20% mortality rate and ICH occurs in 50% [21]. Among our 14 cases diagnosed or suspected as VKDB, five cases (35.7%) had evidence of ICH found either through CT scan or autopsy, and eight cases (57%) resulted in death.

VK is an inexpensive drug, is effective and safe [14], and recommended by WHO as an essential newborn care approach [9]. We strongly recommend introduction of VK prophylaxis at birth in the national policy of essential newborn care. Further, all infants presenting with bleeding, including following vaccination, should be urgently evaluated. Haemoglobin and available coagulation tests should be conducted. Parenteral VK should be administered promptly (after preliminary blood tests) and blood transfusion should be arranged if needed. In many facilities in developing country setting, even tests such as PT/INR may not be available. Therefore, VK may have to be given only on clinical grounds if required. Nevertheless, it should be noted that regardless of how quickly VK is given at presentation, long-term neurological sequelae are common in VKDB with ICH [22]. Therefore, for public health purposes, introduction of VK prophylaxis at birth becomes imperative.

It is already 75 years since a Swedish study [23] showed that VK at birth had five fold reduction in the risk of bleeding to death during the first week of life. American Academy of Paediatrics recommended VK prophylaxis at birth in 1961 [10]. Several other expert professional bodies including the Canadian Paediatric Society [24], [25] have given this recommendation. Prophylactic VK at birth is recommended by WHO [11]. A recent online survey with information gleaned from respondents from about 17 countries (mostly LMIC) shows that the majority of countries had a policy of administration of VK at birth, however, it was not widely practiced in several countries [26]. In countries without such national policy, like Nepal, only some children delivered in the private sector may get this affordable life-saving intervention, which is not equitable. A tertiary referral centre in Nepal identified 16 cases of ICH due to VKDB, of which only 3 (18.7%) had received VK at birth [27]. In current series, we could be certain about only one out of 16 (6%) infants of having received VK at birth.

A Cochrane systematic review has shown that a single dose (1.0 mg) of intramuscular VK given at birth is effective in prevention of classic VKDB; either intramuscular or oral (1.0 mg) VK prophylaxis improves biochemical indices of coagulation status at 1–7 days [6]. However, with respect to effect on late VKDB, neither oral not intramuscular VK has been studied in randomized trials. From observational studies, effectiveness of intramuscular VK prophylaxis in preventing late onset VKDB compared to no treatment or oral dose at birth is evident [15], [16], [28], [29], [30], [31]. Although there were concerns in the past that there may be possible causal relation between intramuscular VK and childhood cancer [32], studies have not revealed any increased risk of intramuscular VK administration and childhood cancer [33]. Therefore, intramuscular VK prophylaxis is proven to be safe and effective for prevention of VKDB.

This case series, with other report of VKDB in Nepal [27], should be taken as an evidence of a cause specific occurrence of infant deaths which can be prevented through a simple and affordable intervention. The relatively low proportion (57.4%) of institutional delivery [2] would be a challenge to effectively deliver a birth dose injection of VK. However, with the government’s policy of increasing the proportion of institutional deliveries, this challenge may be diminished in near future. Between 1990 and 2014, the under-five mortality in Nepal declined sharply by 73% (from 142 to 38 per 1000 live births) [34]. However, the proportion of the neonatal deaths to child deaths increased from 37% to 61% during the same time. Much of the decline in infant and under-five mortality was due to a reduction in post neonatal deaths brought about by several interventions including high infant immunization coverage [35]. Introduction of VK prophylaxis at birth will help in preventing VKDB and could contribute towards reduction of neonatal and early-infant mortality in Nepal.

Based on our observation, we believe that bleeding following vaccination, although coincidental to vaccination, should be recognized as an important AEFI. Brighton Collaboration standardized case definition and guidelines for VKDB (and/or bleeding following vaccination) should be developed. We recommend that VKDB should be suspected in any newborn or young infant presenting with any form of bleeding (either internal or external) under any circumstances including following vaccination. Initially the signs and symptoms of VKDB may be mild but may rapidly progress in severity including to ICH. Therefore, initial symptoms such as mild cutaneous bleeding (or mild but continuous injection site bleeding) or initial symptoms denoting ICH should prompt health care providers to suspect VKDB. As prompt interventions can be life-saving, education of caregivers to promptly seek medical care in the event of these symptoms should be promoted. It is important that health care providers as well as public be made aware of the varied presentation of VKDB [22].

In conclusion, we believe that bleeding following vaccination, especially VKDB, should be recognized as an important AEFI, especially in countries where vitamin K prophylaxis at birth is not given universally. This work will help sensitize service providers to promptly identify and properly manage bleeding cases following vaccination. The available evidence shows that policy and practice of universal vitamin K prophylaxis at birth should be introduced in countries, including Nepal, where it is not yet introduced.

Based on these findings and recommendation from National AEFI Investigation Committee, introduction of universal vitamin K prophylaxis at birth is currently being planned in Nepal.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. NA is the Chair of the National AEFI Investigation Committee (AEFI-IC). RP and ASB are WHO staff and are involved in providing technical assistance for field investigation of serious AEFI cases, and technical and secretarial assistance to the AEFI-IC. BST and JSG are government staff. JSG is member-secretary of the AEFI-IC.