Acute Renal Failure Associated with the Use of Traditional Folk Remedies in South Africa

Abstract

The use of folk remedies is widespread throughout Africa. Acute renal failure (ARF) is one of the most severe, but underrecognized, complications of folk remedy use. This report aims to describe the clinical presentation, outcomes, and nature of renal injury in patients with folk-remedy-associated ARF. Clinical data were evaluated retrospectively in 78 patients with ARF associated with recent folk remedy use. ARF was defined as elevated serum urea and creatinine above the age-appropriate normal ranges, persistent oligoanuria, worsening renal function with time, or need for dialysis. Overall mortality in patients with ARF was 41%. Mortality was higher in adults (45.5%) than in infants (36.6%), in patients with both renal and liver dysfunction (62.5%) than in those with renal dysfunction alone (22.6%), and in HIV-positive (44.4%) versus HIV-negative (34.6%) patients. Vomiting (51.3%) and diarrhea (43.6%) were the most frequent presenting symptoms. Metabolic acidosis (80.8%) and volume depletion (62.8%) were the most frequent clinical findings. The definable causes of ARF were pre-renal (26.9%), acute tubular necrosis (ATN; 26.9%), hepatorenal syndrome (6.4%), urinary tract infection/sepsis (7.7%), and primary renal disorders (7.7%). Twenty-seven patients had concomitant medical conditions unlikely primarily related to folk remedy ingestion. In conclusion, ARF occurring after use of folk remedies in South Africa is associated with significant morbidity and mortality. The most common contributors to ARF in this setting are volume depletion and ATN. Significantly, although a proportion of patients have underlying systemic or renal conditions that may contribute to renal dysfunction, in the majority of patients, folk remedy use appears to be the most likely proximate cause. In view of the large numbers of Africans living abroad, more widespread awareness of this important clinical problem needs to be raised.

• Renal dysfunction
• Acute tubular necrosis
• Dehydration
• Folk remedies
• Metabolic acidosis

Introduction

Acute renal failure (ARF) is a frequent cause of morbidity and mortality in the hospitalized population worldwide, but studies from Africa reveal a strikingly different spectrum of disease from that in the first world. The most frequent causes of ARF in the developed world are therapeutic drug toxicity and renal ischemia resulting from hypoperfusion in cases of sepsis, severe heart failure, radiocontrast administration, especially in the elderly.[1-4] In Africa, in contrast, patients tend to be young, and the most frequent causes of ARF are infections, hemolysis (in areas where G6PD deficiency isprevalent), trauma, herbal toxins, malignant hypertension, and obstetric complications.[5-15] The most frequent infections causing ARF are malaria, typhoid fever, poststreptococcal glomerulonephritis (especially in children), and enteric pathogens causing volume depletion.[5&6], [8&9] The use of folk remedies has been implicated in up to 35% of cases of ARF in Africa.[5&6], [9-11], [13], [16-18] This is likely to be an underestimate of the true incidence because of the difficulty in reliably ascertaining use of folk remedies. Few studies have examined the clinical correlates of ARF associated with the use of African folk remedies. Furthermore, identities of most of the nephrotoxic compounds present in African folk remedies are still unknown. Some nephrotoxic substances that we, and others, have thus far identified from patient urine samples or remedies themselves are Callilepis laureola,[19-21] Securidacea longepedunculata,[7], [21] potassium dichromate,[22] and Cape Aloes.[23] The importance of identifying potential nephrotoxic components in folk remedies has recently been highlighted by the identification of aristolochic acid as the cause of Chinese Herb nephropathy.[24&25] Interestingly, in the animal model of this disease, salt depletion is necessary for the development of renal interstitial fibrosis, highlighting the deleterious role of volume depletion in the setting of nephrotoxic substances.[25]

In this case series, we report the clinical findings and outcomes in 78 adult and pediatric patients who presented to hospital with renal dysfunction associated with recent use of a folk remedy.

Patients and Methods

Ethical clearance for the collection of data and urine samples for analysis of suspected folk remedies was obtained from the ethics committee of the University of the Witwaterstrand.

Patients

Physicians from pediatric and internal medicine wards in the major hospitals affiliated with the University of the Witwaterstrand and some small government hospitals in the surrounding area were requested to notify the investigators of any patients with confirmed use of a folk remedy. Patients giving a positive history of recent folk remedy use, and signing informed consent, were included in this case series. Because of the frequency of use of alternative therapies in the community, many physicians routinely ask whether patients use folk remedies. In the hospital setting, however, there is much stigma attached to the use of folk remedies, and patients are hesitant to admit use. A patient's admission of folk remedy use is, therefore, highly likely to be true, and was taken as confirmation of use. Conversely, denial of folk remedy use is far less likely to be accurate, and therefore, it would not be possible to have a reliable control group of patients with which to compare outcomes or to base prevalence data.

Physicians were requested to submit a brief case report form on each patient, including admission clinical and laboratory data, an early urine sample, and where available, a sample of the traditional remedy. Where possible, full copies of patient records were subsequently obtained from the relevant hospitals. All clinical data available on each patient were analyzed. Laboratory tests were performed in the clinical laboratories of the South African Institute for Medical Research, using standard techniques.

Definitions

ARF was defined as elevations in serum urea and creatinine above the age-appropriate upper limit of normal [urea > 12 mmol/L (34 mg/dL); creatinine > 40 µmol/L (0.45 mg/dL) in children under 1 year, > 80 µmol/L (0.9 mg/dL) in children 1–5 years, > 120 µmol/L (1.36 mg/dL) in patients > 5 years). Severe ARF was defined as a serum creatinine greater than twice the upper limit of normal, persistent oligoanuria, worsening renal dysfunction with time, or the need for acute dialysis. Any patient in whom there was a suggestion of chronic renal failure (e.g., known prior renal dysfunction, small or structurally abnormal kidneys on ultrasound, hypertensive or diabetic retinopathy) was excluded. Severe liver dysfunction was defined as an elevation in transaminases > 200 U/L; total bilirubin > 30 µmol/L, or INR > 1.7. Serum albumin was excluded as an indicator of liver dysfunction, because a significant number of patients were malnourished, had intercurrent infections, proteinuria, or other causes of hypoalbuminemia besides liver dysfunction. Metabolic acidosis was defined as a serum bicarbonate ≤ 18 mmol/L or increased anion gap > 20. HIV tests were carried out with patient or guardian consent. Urine analysis involved the use of dipsticks and urine microscopy. Clinical management and laboratory testing were conducted entirely at the discretion of the patients' physicians.

Results

Clinical and laboratory data were obtained from 127 patients collected over an 18 month period, all withsuspected or confirmed recent use of a traditional remedy. Of the 127 patients, there were 10 in whom strong suspicion of traditional remedy use could not be confirmed on history and 14 for whom clinical data were incomplete; these patients were excluded from further analysis. Of the remaining 103 patients, 78 (76%) with confirmed use of a folk remedy satisfied the criteria for renal dysfunction. Patients were stratified into four separate age groups, as outlined in Table 1.

Table 1 Outcome in Patients with Renal Dysfunction.

	Survived (n = 46)	Died (n = 32)
Male (n = 50)	30 (60%)	20 (40%)
Female (n = 28)	16 (57.1%)	12 (42.9%)
Age:
< 1 year (n = 41)	26 (63.3%)	15 (36.6%)
1–5 years (n = 3)	2 (66.7%)	1 (33.3%)
5–15 years (n = 1)	0 (0%)	1 (100%)
> 15 years (n = 33)	18 (54.5%)	15 (45.5%)
HIV:
Positive (n = 27)	15 (55.6%)	12 (44.4%)
Negative (n = 26)	17 (65.4%)	9 (34.6%)
Unknown (n = 25)	15 (60%)	10 (40%)
Renal dysfunction (n = 78)	46 (59%)	32 (41%)
Severe (n = 51, 65.4%)	30 (58.8%)	21 (41.2%)
Without liver dysfunction (n = 31)a	24 (77.4%)	7 (22.6%)
Renal + liver dysfunction (n = 32)a	12 (37.5%)	20 (62.5%)

Percentages expressed of the total population (n) in the left-hand column.

^a63 patients with renal dysfunction had liver function tests measured.

In the 78 patients with ARF, data on the time of use of the traditional remedy prior to admission was documented for 24 patients. The median time of last use prior to presentation to hospital was 4.5 days (0–30 days). Three patients reported use of remedies on more than one occasion: in one patient, twice daily for 2 weeks; in a second, three times daily for 1 week; the third had taken two different remedies, one 21 days prior, which had made him feel ill, and another (obtained from a different source) 5 days before presentation to counteract the effects of the first. In all 78 patients, remedies were taken orally in 47.4%, as an enema in 10.3%, and by both routes in 12.8%. Two patients (2.6%) inhaled the remedy, and in 26.9% the route of administration was unknown. Twenty-seven patients (32%) were found to have at least one concomitant medical condition, which was unlikely to have been the result of recent toxin ingestion (Table 2). The symptoms of these conditions may have been the original reason for the consultation with the traditional healer, and therefore, the folk remedy could not be implicated alone as the cause of illness or ARF.

Table 2 Coexisting Clinical Disorders.

Clinical disorder	Number of patients
Infections
Blood cultures positive (E. coli, S. pneumonia, E. cloacae)	4
Urinary tract infections	6
Meningitis	2
Septic burn	1
Tuberculosis	1
Falciparum malaria	1
Hepatitis B surface Ag positive	1
Abdominal pathology
Obstructive jaundice	1
Psychosis	1
Pregnancy/intrauterine fetal death	1
Uncontrolled diabetes/hypertension	3
Renal disease
Hemolytic uremic syndrome	1
Membranoproliferative glomerulonephritis	1
Postinfectious glomerulonephritis	3
Systemic lupus erythematosus	1
Childhood
Prematurity	1
Cerebral palsy	1

Sixty-five percent of the 78 patients presented with severe renal dysfunction (Table 1). The severity of renal dysfunction did not affect patient outcome, with an overall mortality of 41%. Duration of hospital stay until date of death was documented in 28 patients, with a median of 3.5 days (range 1–30 days). In the 46 survivors, the median time from admission to discharge from hospital was available in 40, with a mean of 8 days (range 1–60 days). Factors analyzed for their effects on patient outcomes are summarized in Table 1. Adult patients tended to have worse outcomes than patients below 1 year of age. Gender did not appreciably affect patient outcome. Patients who had hepatic together with renal dysfunction had the worst outcome (62.5% mortality), whereas in patients with renal dysfunction alone, the mortality was markedly lower (22.6%). In the 53 patients in whom HIV tests were recorded, 27 (50.9%) were HIV positive, and 26 (49.1%) were HIV negative. This is similar to the rate of HIV positivity in patients being admitted to general medical wards in our hospitals at present. Mortality was higher in patients who were HIV positive (44.4%) than in those who were HIV negative (34.6%).

Clinical Features

The clinical features associated with ARF are shown in Tables 3 and 4. Prior to presentation, the most frequent symptoms were vomiting (51.3%) and diarrhea (43.6%). At the time of presentation, 53 patients (67.9%) had shock or dehydration: dehydration was at least 5% in 33 (42.3%), and 23 patients (29.5%) were in hypovolemic shock. By contrast, 10% of patients were volume overloaded on admission. In the first 24 h following admission, the most frequent clinical findings were altered mental status (seizures, coma, encephalopathy: 43.6%), Kussmaul breathing (28.2%), oliguria (26.9%), respiratory distress (16.7%), and fever (14.1%).

Table 3 Clinical Features in Patients with Renal Dysfunction.

	Present study (n = 78)	Prior studies	Reference
Metabolic acidosis	63 (80.8%)	62.5–86%	[29], [34]
Dehydration	49 (62.8%)	42.9%	[34]
Vomiting	40 (51.3%)	37.5–83.3%	[29&30], [34]
Jaundice/liver failure	32 (50.8%)	26.4–57%	[10], [30], [34]
Seizures/encephalopathy	34 (43.6%)	15.4–83.3%	[29&30], [34]
Diarrhea	34 (43.6%)	25–50%	[29&30], [34]
Shock	23 (29.5%)	4–87.5%	[29], [34]
Kussmaul breathing	22 (28.2%)	18.8–33.3%	[30], [34]
Respiratory distress/cough	13 (16.7%)	19.8%	[34]
Fever	11 (14.1%)	25–67%	[29&30])

Table 4 Features Associated with Renal Dysfunction.

Clinical finding	Number of patients (%)
Volume status
Dehydration and shock	53 (67.9%)
Dehydration	49 (62.8%)
Shock	23 (29.5%)
Volume overload	8 (10.3%)
Metabolic acidosis	63 (80.8%)
Increased anion gap	27 (34.6%)
Normal anion gap	36 (46.2%)
Mixed	14 (17.9%)
Electrolytes (mmol/L)
Hypernatremia (Na ≥ 150)	12 (15.4%)
Hyponatremia (Na ≤ 130)	24 (30.8%)
Hyperkalemia (K ≥ 6)	10 (12.8%)
Hypokalemia (K ≤ 3)	15 (19.2%)
Urinalysis (n = 60)
Proteinuria	41 (68.3%)
Hematuria	33 (55%)
Urine microscopy (n = 22)
Granular casts	19 (86.4%)
Red cell casts	4 (18.2%)
White cell casts	2 (9.1%)
Hyaline casts	3 (13.6%)
No casts	2 (9.1%)

Percentages expressed of total population with renal dysfunction or of (n) in left-hand column.

Electrolyte abnormalities were common (Table 4). In the 24 patients who were hyponatremic ([Na⁺] < 130 mmol/L) the majority were volume depleted. Two patients died of hyperkalemia soon after admission, with potassiums of 9 and 10.3 mmol/L, respectively. Metabolic acidosis was present in 63 (80.8%) patients, 42.9% of whom had a high anion gap, 57.1% a normal-anion gap, and 22.2% a mixed metabolic acidosis. In all patients with high anion gap acidosis, the magnitude of the gap appeared to be appropriate for the degree of renal or hepatic dysfunction or shock, except for one, with both severe hepatic and renal failure, in whom the anion gap was 70. This patient died before other sources of unmeasured anions could be investigated.

Serum albumin was measured in 65 patients with a median of 30.1 g/L (range 9–46 g/L). Albumin was below the lower limit of the normal range (35 g/L) in 46 patients (70.8%). Proteinuria was present in 41 (68.3%) patients (Table 4), with 10 patients having more than 3 proteinuria on dipstick. Five of the 10 patients had no obvious cause of proteinuria other than folk remedy use, whereas the other five were individual patients with a positive antistreptolysin O (ASO) titer, membranoproliferative glomerulonephritis (MPGN), diabetes mellitus, gross hematuria, and malaria. Hematuria was present on dipstick in 55% (Table 4), one patient had gross hematuria. Urine microscopy findings are listed in Table 4.

The clinical course of renal dysfunction was documented in 57 patients. Renal function improved with conservative management in 36 patients (63.2%), with a median time to recovery of 3.5 days (1–21 days). Renal function worsened in 21 (36.8%) patients, 16 of whomdied. Dialysis was available only to adult patients in one of the participating hospitals. Ten of the 23 adult patients included from this hospital required hemodialysis. Indications for dialysis were life-threatening hyperkalemia,[3] persistent oliguria,[5] uremic encephalopathy,[1] and severe refractory acidosis.[1] Five patients died while still receiving hemodialysis after a median of 11 days from admission (5–21 days). In the surviving five patients, four came off dialysis after 52, 25, 23, and 22 days, respectively, and one, in whom renal biopsy confirmed hemolytic uremic syndrome (HUS) with significant interstitial fibrosis, remains on chronic dialysis.

Pathophysiology of Acute Renal Failure

Identifiable factors contributing to ARF in all 78 patients are listed in Table 5. All of these factors, with the possible exceptions of urinary tract infections (UTI) and intrinsic renal disorders, are most likely the result of folk remedy ingestion either through direct nephrotoxicity, or indirectly through volume depletion. In 21 (26.9%) patients, recovery of renal function occurred within 3 days after volume resuscitation; these patients are presumed to have had pre-renal renal dysfunction. Acute tubular necrosis (ATN) was confirmed in 21 (26.9%) patients by the presence of granular casts on urine microscopy,[19] high fractional excretion of sodium,[3] or by renal biopsy or autopsy.[4] Histologically, in two patients, biopsy confirmed features of ATN in addition to HUS and MPGN, respectively, and in two further patients, postmortem examination reported ATN in one and “shock kidney” in the other. Three patients had elevated ASO titers, suggesting possible postinfectious glomerulonephritis. One patient had acute systemic lupus erythematosus (SLE) confirmed by positive serology. Six patients were found to have UTIs, two with urinary white cell casts and one with a functional bladder abnormality on voiding cystourethrogram (VCU). Five patients had hepatorenal syndrome (HRS) with deteriorating oliguric renal function as liver function worsened. In one of these patients, renal function later improved, commensurate with liver function. A further four patients, who were oliguric with severe liver dysfunction, died soon after admission, before a diagnosis of HRS could be made. In the remaining patients, the exact nature of renal injury was unclear.

Table 5 Factors Contributing to Renal Dysfunction.

Cause	Number of patients
Hypoperfusion (shock + dehydration)	53
Prerenal	21
ATN	21
Hepatorenal syndrome	5
Urinary tract infection/sepsis	6
Intrinsic renal disorders
MPGN/nephrotic syndrome	1
HUS	1
Postinfectious glomerulonephritis	3
SLE	1

Discussion

Systematic study of patients using traditional remedies in Africa is not straightforward. Many patients resort to Western medicine only after folk treatment fails or a complication ensues. When patients do present to hospitals, there is much secrecy surrounding use of folk remedies because of fear of stigmatization or social pressures. In addition, the practical challenges of day-to-day hospital practice in South Africa make enrollment of eligible patients and collection of clinical data difficult. These facts make it almost impossible to obtain accurate statistics as to the frequency of use of folk remedies in the hospitalized population and to determine their true impact on morbidity and mortality in our patients. We do believe, however, that this is an important clinical problem to tackle, because in addition to significant patient morbidity, the impact of folk remedy toxicity leads to increased costs to our already strained health care system. Furthermore, as use of alternative therapies is increasing in the Western world, complications arising from use of these remedies will become a more global problem.

The use of folk remedies in daily life is widespread in Southern Africa, with reported rates of up to 80% in the indigenous Black communities.[26-28] Reasons for consultation of traditional healers include to protect or strengthen the child, to increase sexual potency or to treat infertility, to procure abortion, to remove evil spirits, to get rich, and as prophylaxis against witchcraft.[7], [29-33] In addition, folk remedies are sought for physical complaints. Given the vast number of people using these remedies, most appear to do so without experiencing adverse effects. In some cases, however, as with Western medicines, “iatrogenic” complications arise, and patients present to hospitals with a variety of complaints.[7], [19], [22], [29&30], [34-37] ARF is one of the most life-threatening of these complications and is associated with significant morbidity and mortality across the continent.[1], [7], [27], [30], [34], [38]

Males predominated in the two major age groups included in the study (approximately 66%). This finding issimilar to the 62% found by Gold in 1980.[34] The reason for this is unclear but may lie in traditional beliefs. Males are the “heads” of households and may have increased access to herbal remedies, parents may feel that male children need more “protection,” and adult males may take remedies for problems specific to their sex, e.g., sexual potency. Most studies where females predominate have described small groups of patients, and the remedies used were usually to induce abortion.[7] Social and cultural differences in different parts of Africa, and differences in source population within hospitals, are all likely to impact gender prevalences in any such study. It is also possible that no true difference between genders exists, but that remedies used by females for different indications than in males may be less toxic, or females in general may be more reluctant to admit use of a folk remedy.

The most frequent clinical findings presented here are similar to those reported in previous studies in which volume depletion and shock were frequent (Table 4). In cases in which folk remedies are sought because of abdominal discomfort, constipation, or to “clean the stomach,”[7], [23], [30], [39] volume depletion, resulting from vomiting and diarrhea, is expected. In fact, many remedies prescribe self-induced vomiting after ingestion or large-volume enemas as part of the catharsis rituals. Volume depletion and hypotension were found to play a significant role in ARF associated with the use of folk remedies in this study and are both factors known to enhance nephrotoxicity of many drugs and compounds.

The majority of our patients presented with a metabolic acidosis (Tables 3 and 4), a frequent finding associated with the use of traditional remedies.[20], [30], [34&35] In the 57% of patients with a normal anion gap acidosis, diarrhea was a common symptom, and the most likely cause. On the background of nephrotoxicity, renal tubular dysfunction may have contributed in some but could not be excluded, because urine electrolytes were not measured. The remaining 43% of patients with metabolic acidosis had an elevated anion gap, which is expected in patients with severe renal and liver dysfunction or shock. We had only one patient in whom the anion gap suggested the possibility of additional unmeasured anions. This is consistent with prior data. In a study of metabolic acidosis associated with folk remedy toxicity, Nkrumah et al. found that renal failure and lactic acidosis accounted for the elevated anion gap in 14 of 20 children with severe metabolic acidosis.[35] They failed to find any evidence that a toxin or a metabolite had made a direct contribution.

The overall mortality rate in this study (41%) is comparable to that reported in other African studies.[10], [29], [34] Mortalities range from 24%–75%, with the highest percentage pertaining to patients in an intensive care unit.[29] Mortalities were similar in males and females. Unexpectedly in this study, adult patients had a higher mortality than patients less than 1 year of age. This difference may be explained by several observations: there was a higher incidence of volume depletion in the pediatric group, in whom renal function was often reversible with rapid volume resuscitation; and there was a higher incidence of severe liver dysfunction in adult patients, a factor found to adversely affect prognosis. As expected, mortality was higher in patients who were HIV positive due to the often advanced stage of disease at clinical presentation, the lack of availability of antiretroviral therapy in state hospitals, and the fact that these patients are more likely to be volume depleted.

Despite its severity on admission, in many cases, renal failure often improved with adequate supportive care. Overall, 26.9% of patients had pre-renal renal dysfunction that recovered within 3 days with intravenous hydration alone, and a further 26.9% had ATN, which also tends to improve given time. The presence of concurrent renal and liver dysfunction, however, was associated with a significantly worse prognosis compared to the presence of renal dysfunction alone. This is similar to findings of Seedat et al., where jaundice was associated with a poorer prognosis in patients with ARF, the majority of whom had used herbal remedies.[10] This observation is not surprising, as supportive care in patients with severe liver dysfunction is more complicated, especially given limited intensive care facilities, making it more difficult to “ride out” the liver failure until recovery occurs.

The exact nature of the renal injury is often difficult to define clinically. ATN was the most frequent identifiable cause of renal dysfunction. The high frequencies of proteinuria and microscopic hematuria, although nonspecific, also argue for intrinsic renal damage. Two patients had unexplained high-grade proteinuria on dipstick with accompanying hypoalbuminemia, but not enough collateral data were available to diagnose nephrotic syndrome. At least five patients had convincing evidence of HRS, likely directly associated with the use of a folk remedy, as other causes of liver failure, e.g., paracetamol overdose, were excluded. Rhabdomyolysis has been implicated as a cause of ARF related to ingestion of a Chinese herbal remedy,[40] but we found no cases in this series.

All 78 of the patients included in this case series were admitted to hospital and gave a positive history of recent use of a folk remedy, in most cases, within the week prior to presentation. This information is useful in trying to establish a cause-and-effect relationship between remedy ingestion and subsequent symptoms or illness. It is impossible, however, to establish direct cause and effect unless one can identify toxic substances in the patient's urine or obtain the remedy and establish its toxicityexperimentally. Evidence is slowly emerging that supports selective toxicities of several known plant remedies: aloesin and aloeresin A, present in Cape aloes; atractyloside, present in Callilepis laureola; and methyl salicylate, present in Securidacea Longepedunculata. These have all been reported to be nephrotoxic; pyrrolizidines, present in Senecio and Crotalaria, have been found to cause hepatic veno-occlusive disease; atractyloside, present in Callilepis laureola, and polycyclic hydrocarbons, present in Euphorbia ingens, have been found to have hepatotoxicity.[7], [19], [20], [23], [27], [41-43]

Renal biopsies have rarely been performed in patients with nephrotoxicity associated with the use of African folk remedies. In one study, the kidneys were described as either “normal” looking or pale with evidence of hypoperfusion at autopsy.[29] In other small studies, features of ATN and interstitial nephritis have been documented.[7], [10], [34] In the present study, only three patients underwent postmortem examination: the reports indicated that one had a “swollen” kidney, one a “shock” kidney, and the third ATN. The patient with shock kidney was found to have atractyloside in her urine, consistent with use of the plant Callilepis laureola.[19] In the current case series, six patients had evidence of intrinsic renal disorders, and a further six had urinary tract infections. It seems, therefore, that a proportion of patients with renal failure “associated” with the use of folk remedies have underlying renal disorders that may account for all, or at least some, of the observed renal dysfunction. It is important to bear in mind, however, that a kidney with preexisting pathology may be more susceptible to the effects of nephrotoxins, and these patients may therefore have been more vulnerable to toxicity from the folk remedies.

Ten patients in this study received hemodialysis. Mortality was high (50%), but in four of the surviving five patients, renal function improved after at least 3 weeks. In an earlier study, Gold reported that peritoneal dialysis did not affect the outcome of patients with folk-remedy-associated ARF.[34] The majority of patients in this study only received a single dialysis, however, and therefore may not have had severe renal dysfunction. In contrast, two studies from Nigeria and Kenya each describe the use of hemodialysis in six patients poisoned with herbal remedies. Both reported renal recovery in more than half of the patients, although time to recovery was prolonged.[17], [30]

Of interest in this study is the documentation of underlying clinical disorders in a proportion of patients, which may well have been the reason why the patient sought the counsel of a traditional healer (Table 2). Consistent with our findings, Nyazema also reported a 30% incidence of “other” disorders in patients using folk remedies.[27] In these patients, it is difficult to determine how and whether the underlying illness contributes to the ARF, and whether the folk remedy can be implicated at all. On the other hand, more importantly, in the remaining majority of patients, no other cause of illness was found besides the history of folk remedy use. Whether the effect on renal function is direct nephrotoxicity or is augmented by renal hypoperfusion from dehydration or hypotension, the circumstantial evidence spanning many years supports a consistent association between folk remedy use and ARF in Africa.[5&6], [9-11], [13], [16-18], [44]

In conclusion, we reported on a series of patients with ARF presenting to hospital after recent use of a traditional folk remedy. Mortality and morbidity were high and in keeping with those described by other authors. Volume depletion and shock were important contributors to the observed ARF. We suggest that it is highly possible that some folk remedies, while not intrinsically toxic in a euvolemic patient, may have enhanced nephrotoxicity in the setting of volume depletion. We, therefore, do not intend to imply that all folk remedies are harmful, and we are aware that their use is widespread and plays a significant role in the lives of most people in our community. We do believe, however, that collaboration between traditional health care practitioners and the medical community is urgently needed to identify potentially harmful compounds that are present in folk remedies and to modify their use in patients at greatest risk of complications. Furthermore, as many Africans now live in Europe and North America, physicians practicing in these countries need to be aware of the widespread use of traditional remedies and need to inquire about their use.

Acknowledgments

The Medical Research Council of South Africa (VS) and the Friedel Sellschop Award from the University of the Witwaterstrand (VS) supported these studies. We are grateful for the cooperation of doctors and patients who participated in the study.