Quality of Phenobarbital Solid-dosage Forms in the Urban Community of Nouakchott (Mauritania)


Address correspondence and reprint requests to Dr. P-M. Preux at Institute of Neuroepidemiology and Tropical Neurology (EA 3174), 2 Rue du Docteur Marcland, 87025 LIMOGES Cedex, France. E-mail: preux@unilim.fr


Summary: Purpose: Epilepsy is a major public-health problem in Africa. The quality of available drugs is a limiting factor for an adequate management. The aim of this study was to describe the proportion of poor-quality phenobarbital (PB) solid-dosage forms and evaluate the factors associated with its quality in Nouakchott (Mauritania).

Methods: A cross-sectional study was carried out within pharmacies, hospitals, and on the parallel market in March 2003. PB samples were bought by a native person and then assayed by a liquid chromatography method. A package was considered to be of good quality if the active-substance average content was between 85 and 115% of the stated content printed on the packet.

Results: Forty-five pharmaceutical stores were visited, enabling us to collect 146 samples of PB. Three brand names were available in Nouakchott. They originated from France, Morocco, Senegal, and Egypt.

Results: A prevalence of 13.7%[95% confidence interval (CI), 8.8–20.0] of poor-quality PB was found. All samples from Morocco were underdosed. The generic active content was satisfactory, but saccharose, an excipient with a potential side effects, was identified. Two factors associated with the good quality of PB have been put forward: tablets manufactured in France and loose packaging as generics conditioned in such a way were of good quality.

Conclusions: This study shows that the quality of antiepileptic drugs in Africa is still worrying. The setting up of medicine quality control in Mauritania is legitimate. Considering the good quality of generic PB and its lower cost, this type of medicine should be promoted in this region.

Epilepsy is a public-health problem in tropical areas because of its high frequency, insufficient treatment (treatment gap), and socioeconomic impact (1,2).

In most developing countries (DCs), Phenobarbital (PB), recommended by the World Health Organization (WHO), remains the most prescribed medicine because of good availability, low cost, and large activity spectrum on different forms of epilepsy, except absences (3). However, the quality of available medicines may be uncertain and could influence the effectiveness and harmlessness of treatments. A large number of examples questioning the quality of drugs illustrate this public-health problem (4,5). Some studies have shown a 10–40% prevalence of substandard medicines according to drugs and countries (6–9).

The aim of this study was to evaluate the prevalence of poor-quality PB in the stores delivering this medicine in Nouakchott (Mauritania) and to determine the factors associated with the poor quality of PB.


A cross-sectional study was carried out during March 2003 in the Nouakchott urban community in Mauritania (West Africa). Patients could buy antiepileptic drugs (AEDs) either in the pharmacies of public hospitals, in 87 private pharmacies, or in 63 pharmaceutical retail stores run by people with no specific training. The private sector got supplies from nine private wholesalers, and the public sector, from the Purchase Centre of Essential and Edible Medicines.

Only oral solid-dosage forms of PB were studied. PB was available on prescription, either in box, blister pack, or per unit, as tablets, under several brand names and active-substance contents. The statistical unit was defined as a PB box (or 10 tablets, if retailed), whatever the brand name or the dosage.

A neurologist or a nurse from the Neuro-Psychiatric Centre (NPC) visited pharmaceutical stores in different Nouakchott areas, after random selection from a list of stores in the neighborhood. The manager or an employee was met and questioned on the store functioning and on storage and supply conditions.

Then another NPC nurse went to the previously visited pharmaceutical stores and bought a PB box or 10 units of the various dosages and specialties available with a prescription.

Another questionnaire was filled in about quality criteria (packaging, presentation, name, laboratory name, country of origin, expiry date, presence of guidelines, language of the guidelines, registration number, price, dosage mentioned on the pack, and tablet appearance).

The study was supplemented by a visit to the different possible retail outlets of the unofficial market, searching for AEDs; the addresses were communicated by word of mouth.

Assays were performed in the Department of Pharmacology-Toxicology (DPT) of Limoges University Hospital through high-performance liquid chromatography (HPLC) with ultraviolet detection (UV), validated for determination of drugs in pharmaceutical preparations (10). The system for HPLC analysis was composed of: a chromatographic pump (Waters 510, Waters, Saint-Quentin-en-Yvelines, France), a Shimadzu SPD6AV UV-detector (Shimadzu, Croissy-Beauborg, France), a C18 column (150 × 4.6 mm i.d., 5 μm). The mobile phase consisted of methanol: KH2PO4 (pH 3; 1 mM) (45:55, vol/vol) at a flow rate of 1.3 ml/min; UV absorbance detection at 220 nm. Tablets were dissolved in ultrapure water, agitated for 15 min, and diluted with the mobile phase. For every package, 10% of the tablets were selected at random and assayed. An average amount of PB per package was calculated and compared with that mentioned on the package. Three result categories were defined: (a) correctly dosed sample: average between 85 and 115%; (b) underdosed sample: average, <85%; and (c) overdosed sample: average, >115%.

The tablet quality study was completed by three standardized pharmaceutical tests (mass uniformity, resistance to rupture, disintegration time) according to specifications from the European Pharmacopoeia, 4th edition, 2002 (11) in the Laboratory of Galenic Pharmacy at the Limoges Faculty of Pharmacy. Mass uniformity study allowed checking that the individual mass of an intake unit remained within a 5% deviation around the average mass of a batch of intake units. Tablet resistance to rupture was measured by the diameter crushing test. The force necessary to bring about rupture was converted into resistance, which gave a general survey of the internal cohesion of tablets. The disintegration test evaluated the tablets' more or less great ability to disintegrate in distilled water within the stipulated time (15 min). These tests permitted the comparison of the characteristics of the PB tablets collected.

Finally, an analysis using Fourier transform infrared spectrometry (FTIR) was performed in the DPT. To get a good transmission, part of a tablet was dispersed in KBr and ground in an agate mortar. The powder was compressed to form a thin pellet, which was placed in a Spectrum One Perkin Elmer (Perkin-Elmer, Courtaboeuf, France), version B-53607 spectrophotometer (frequency range, 7800/350 cm; resolution, 0.5/cm−1). The FTIR permitted the comparison of the different samples spectra, and with a bank of spectra, to identify the excipients.

Statistical analysis was carried out by using Statview 5.0 (SAS Institute, Cary, SC, U.S.A.) and SAS 8.0 software (SAS Institute). After a descriptive analysis of the variables, a multivariate logistic regression was performed to determine the independent factors linked to the quality of PB. The Hosmer and Lemeshow method was used with a step-by-step descending procedure. Selection of variables was made from the statistical association with the dependent variable (quality of samples) in a bivariate analysis (p ≤ 0.25). The significance level was fixed at 5%.


In total, 35 (77.7%) private pharmacies and 10 (22.3%) retail pharmacies were visited. AED boxes were stored on shelves (84.4%) in the shops. They were sheltered from the sun (88.9%) and humidity (86.7%) but often exposed to dust (68.9%). Restocking with AEDs at wholesalers was done every week (82.2%), suggesting that little stock was contained in these stores. Supply could also be received (52.8%) from the parallel market or from abroad (Senegal, Tunisia).

Characteristics of PB samples collected and results of analysis are reported in Table 1. Generic PB was cheaper than Gardenal (1.4 times for 50 mg, 2.3 times for 100 mg). PB was not found on the parallel market.

Table 1. Characteristics and results of analyses on active principle content analyses of phenobarbital samples collected in Nouakchott (Mauritania), March 2003

(Country of origin)

Number of

of expiry

Indication of

Good quality
Correctly dosed
Poor quality
  1. Every sample consists in 10% of a packaging.

  2. Limit values of dosages for a correctly dosed sample (dosage indicated on the package ± 15%):

  3.  For 50 mg: 42.5 to 57.5 mg/L

  4.  For 60 mg: 51.0 to 69.0 mg/L

  5.  For 100 mg: 85.0 to 115.0 mg/L

  6. The values of dosages for samples considered as poor quality are indicated in square brackets [mean±SD].

Gardenal® 5035 (boxes)YesYes3212
 (France) [39.9±1.3 mg/L][59.6±1.6 to 61.6±13.2 mg/L]
Gardenal® 5010 (boxes)YesNo0100
 (Morocco) [39.2±2.5 to 41.6±0.4 mg/L] 
Phenobarbital Generic 501 (tube)YesNo100
Phenobarbital Generic 501 (per unit)NoNo010
 (Senegal) [38.6±1.6 mg/L] 
Sominal® 6030 (blister packs)YesNo2550
 (Egypt) [46.9±0.1 to 48.3±4.4 mg/L] 
Gardenal® 10036 (boxes)YesYes3510
 (France) [84.4±12.7 mg/L] 
Phenobarbital generic 10033 (per unit)NoNo3300
Total146 126182

Sample quality was rated as poor in 13.7% (CI, 8.8–20.0) of medications.

The mass of French 100-mg Gardenal tablets was smaller (177 ± 2 mg/195 ± 6 mg), disaggregation time was shorter (180 s/600 s), and resistance to rupture higher (3.215 ± 0.284 MPa/1.099 ± 0.441 MPa) than those measured with 100-mg generic PB, indicating that qualitative as well as quantitative compositions of these two drugs were different (Table 2). FTIR analysis identified saccharose (excipient with potential side effects) in generics. Similarly, Sominal contained lactose (excipient with negative side effects).

Table 2. Results of pharmaceutical tests (mass uniformity, resistance to rupture, disintegration time) of phenobarbital samples collected in Nouakchott (Mauritania), March 2003
Limits of average
mass (mg)
Resistance to
rupture (MPa)
time (sec)
  1. Mean ± SD

Gardenal® 506.870 ± 0.0021.678 ± 0.00488.2 ± 0.9 79.4 – 97.1Yes3.598 ± 0.263120
Gardenal® 506.865 ± 0.0051.620 ± 0.03386.5 ± 2.2 77.8 – 95.2Yes2.998 ± 0.231 90
Phenobarbital Generic 505.366 ± 0.0072.973 ± 0.03976.2 ± 1.1 68.6 – 83.8Yes2.566 ± 0.263 60
 Sominal® 606.870 ± 0.0232.952 ± 0.051141.1 ± 3.9  126.9 – 155.2Yes1.970 ± 0.482 30
Gardenal® 1008.936 ± 0.0082.506 ± 0.037177.2 ± 1.3  163.9 – 190.5Yes3.215 ± 0.284180
Phenobarbital generic 1007.892 ± 0.0133.012 ± 0.101195.2 ± 6.2  180.6 – 209.9Yes1.099 ± 0.441600

Two independent factors were linked to PB quality: the country where the drug was manufactured and the nature of the drug (either proprietary or generic). Tablets manufactured in any country but France were more likely to be of poor quality (OR, 9.7; CI, 2.9–31.8). Loose packaging was associated with a smaller risk of encountering poor-quality PB (OR, 0.05; CI, 0.01–0.42). No factor linked to the structure itself showed a significant risk of purchasing poor-quality PB.


AEDs in DCs do not escape poor-quality problems. The prevalence found here is comparable to that in other studies on the quality of medicines concerning mainly antiinfective drugs (6–9). A low quality of PB originating from Morocco was identified, a peculiarity known by some patients, who sometimes continued to buy this drug because of its lower cost. A WHO report warned against the poor-quality medicines from Morocco on account of poor technology, poor controls, and weakness of the administration with respect to enforcing regulations (12).

In studies on drug quality, sampling methods are not often clearly explained. We contemplated various sampling possibilities. A cluster sampling of pharmaceutical stores could have been planned after a list of stores had been established in every Nouakchott neighbourhood (no list available). In each selected store within this cluster sampling, it would have been necessary to take all available PB samples, which would have been expensive, unethical, and detrimental to the population. Another possibility would have been to perform a sampling of PB in all the pharmaceutical stores of Nouakchott. This was difficult to consider because of insufficient time and means. So it was decided to visit only part of the pharmaceutical stores, selected at random, and to collect only one sample of PB in every dosage form and brand name. The purchasing of samples with a prescription by a native permitted us to collect the medicines that were usually sold to the population.

The definition of PB sample quality in this study was more restrictive than in others on this topic (7,8) as it was based on the result of the assay of the active constituent in a unit of drug intake, which was to be between 85 and 115% of the content printed on the package. The definition of the quality based on the European pharmacopoeia (11) is more oriented toward checking the homogeneity of a batch than toward interbatch similarity. Consequently, if all unit samples were underdosed, they would lie within the limits of the average content of a batch and thus would be considered of good quality. But considering the dosage printed on the package and setting beforehand of limits below and over this value, an overdosage or underdosage of the tablets assayed from the sample can be shown. This is essential so that a patient could take a proper dose of an AED for good control of seizures.

However, the approach to the quality of a medicine based on the assay of its active constituent content is partial. Infrared spectrometry and galenic tests identified two excipients with potential side effects, lactose and saccharose, which are contraindicated in people with lactase or saccharase enzyme deficits. The frequency of these enzyme disorders is high in blacks (13), which suggests a possible bad tolerance of medicines containing lactose or saccharose among African populations.

Bioavailability is determined by numerous factors: tablet composition, nutritional state, feeding behavior, individual physiologic variations, and drug association. A study on the pharmacokinetics of carbamazepine showed that two generic drugs were not bioequivalent to Tegretol (14). PB bioavailability and dissolution tests were not performed, as our aim was to study the pharmaceutical quality of the drug.

In our study, the absence of PB on the parallel market was encouraging. As data on medicine quality or on the supply circuit in the parallel market are lacking, it is impossible to put forward a hypothesis based on a higher risk of finding poor-quality medicines on this market.

Generic PB has received interest in Mali for at-home treatment of epilepsy, demonstrating the effectiveness on seizures, the paucity of side effects, and the low cost ensuing (15).

On the whole, good-quality generic PB, which remains less expensive than brand-name drugs, should be promoted in Mauritania for a better control of epilepsy. A regular surveillance will have to be achieved and maintained, for these products are not immune from a quality defect some day. The setting up before long of a control structure of medicine quality in Mauritania is wholly legitimate. Analyses on the quality of other AEDs widely used in Mauritania (carbamazepine, sodium valproate) must be continued, including bioequivalence tests.


Acknowledgment:  This study was financially supported by the Technologic Science Health Doctoral School of the University of Limoges, by the Conseil Régional du Limousin, and by the Institute of Neuroepidemiology and Tropical Neurology (EA3174). We thank those who helped us to carry out this study: the staff of the Neuro-Psychiatric Centre of Nouakchott, the staff of the Department of Pharmacology-Toxicology of the Limoges University Hospital (Prof. Lachatre, Prof. Marquet, and Prof. Moesch), and the staff of the Laboratory of Galenic Pharmacy of the Faculty of Pharmacy of Limoges (Prof. Chulia) and Dr. Ratsimbazafy.